The Fauna and Éeographfy of the M5 Dive and Lacca'dive ...

$6e I ] , qf ValuMe wZYZfie’

pué/zbéed'm

TheFauna and é eographfyof the

M 5 diveand Lacca'

diveArchipelagoes

Being t heAecount of thew ork"carried ”

on and

of theCollec t ions madeby'

anExpedition

duringi

t

'

he'

years'

1 899 and”

1900

Edi ted by

j. STA Y

r'

GARDINER, M .A.

Fellow of G onville and Caius College"

and Balfour Studentof t heUni versit y of Cambridge.

v

ogifie PART

PEWit h Plates

i

I

I

“

and Tex t—Illustrat ions t .

t iAM BRIDGE'

:

a t theUnivers i ty Press.

LONDON z‘

C.J. CLAY AND S ONS,CambridgeUnivers ity Press W arehouse,

The Fauna and Geographyof t he

M aldive and LaccadiveA rehipeagoes

VOLUME I. PART

1 01111011 : 0. J . CLAY AND SONS ,

CAMBRIDGE UNIVERSITY PRESS W AREHOUSE,

AVE MARIA LANE ,AND

H. K. LEW IS,136 , GOW ER STREET , W .C.

fi lasgom: 50, W ELLINGTON STREET.i teippig : F.A.BROCKHAUS.

{Rein y ork : THE MACM ILLAN COMPANY.3301111189 : E. SEYMOUR HALE.

[All R igh ts reserved ]

TheFauna and Geographyof t he

diveand LaccadiveArch ipeagoes

Being t heAccoun t of t heW ork c arried on and

of t heCollec t ions madeby an Ex ped ition

during t heyea rs 1 8 99 and 1900

Edi ted by

J. STANLEY G ARDINER,M .A.

Fellow of G onville and Caius College and Balfour St udent

of t heUn ivers it y of Cambridge.

VOLUM E I. PART I.

W it h Plates I—V and Tex t - Illus t rat ions 1—23

CAM BRIDG E :

a t theUn ivers i ty Pres s.

PRINTED BY J . AND C. F. CLAY ,

AT THE UNIVERS ITY PRESS.

PRE FACE .

IN March , 1899, I left England , in pursuance of my appoin tment as Balfour Studentof the Universi ty of Cambridge , w i th a commission to explore and investigat e the CoralReefs of the Laccadives , Maldives , and Ceylon. As t he sou t h - west monsoon is unsuitable forwork from smal l craft in the Indian Ocean , i t was proposed t o devote t he summer of 1899to a thorough survey of t he atol l of M in ikoi

,the most southern reef of the Laccadives.

In regard t o subsequent work the Managers of t he Balfour S t udentship gave me completelat itude t o do as migh t seem fit .

Subsequent donations of £300 from t he G overnment G rant , administered by t he RoyalSociety

,and of £30 from the B ritish Association

,decided me to proceed in October t o the

Maldive Archipelago for an extended winter cruise. The lat ter was ent irely successful inits main resu lts

,al though the e x pedi t ion w as considerably curtailed and seriously crippled

owing to t hegeneral rise in prices caused by t hewar in South Africa.

I would like here to express t he indebtedness of my part y to His Excellency SirE. Noel Walker, lat e Acting - G overnor of Ceylon ; to t he Hon. Mr Taylor,lat e Act ing - Colonial Secre t ary ; t o H is H ighness Mohammadu Imaduddin ,

Sultan of t he

Maldives,and his Viziers ; to the B oard of Trade for passages to and fro betw een Ceylon

and M inikoi ; to Capt . Channer, R.N.,formerly an officer of H.M.S. Challenger, now Super

intendent of Ligh t houses in t he Ceylon dis t rict , for many valuable hint s and much localinformation ; t o Sheikh Jeevunjee Noorbhai for his generous help and assist ance during t he

time we were in the Maldives ; t o Rear - Admiral Sir W. Whart on , Hydrographer, for adviceand loan of scien t ific instrument s ; t o Dr David Sharp for t aking charge of t he insectcollections ; to Sir John Murray

, Prof. Judd and Prof. Agassiz for advice andassist ance in many ways ; and indeed t o al l kind friends and con t ributors for t heir aid.

Above al l I wish to acknowledge t he great obligations of mysel f and my party t o

Mr Adam Sedgwick for his great assistance and very active in t erest in t he progress andwork of t he expedi t ion.

In t he work I was volunt ari ly assisted by Mr L. A. Borrada ile (Selwyn Col lege) andMr C. Forster Cooper (Trini t y College), who have largely cont ributed t o what ever measureof success may have been at t ained. During five weeks in March and April , 1900,

Mr Fors ter Cooper t ook charge of the work single - handed a t a time when I was incapacita ted by fever. My thanks are also due to Capt . Molony , S S. Ileafaee, for his helpin surveying Suvadiva and Addu at ol ls ; and furt her t o Mr Geo. Sheldrak e , in charge ofS S .Thrunscoe, during our visi t st randed on M inikoi , for pract ical help on many occasions.

CONTE NTS OF VOL . I . PART I .

Report s.PAGE

1. Int roduc t ion : Na r ra t i ve and Rou te of t he E x pedit ion. W it h

Tex t - F igs. 1 and 2

By J . STANLE" GARDINER, M A.

2. The Maldi ve and Laccadi ve G roups , w ith Notes on ot her Coral

Format ions in the Indian Ocean. W it h Pla tes I and II,Tex t

F igs. 3—1 1By J . STANLEY G ARDINER, MA .

3. Hymenoptera

By P. CAM ERON.

4. Land C rus t aceans. W i th Pla te III,Tex t - F igs. 12—23

By L. A. BORRADAILE,M.A., Lect urer in Nat ura l S ciences of Selwyn College,

Cambridge.

5 . Nemerteans. W i th Pla tes IV and V

By R.C. PUNNETT, B.A.

INTRODUCTION.

NARRATIVE AND ROUTE OF THE EXPEDITION.

(W i t h 2 Maps.)

AT t he end of March , 1899, I left England for Ceylon in the company of Mr L. A.Borradaile. On arrival at Colombo we found t hat the Board of Trade s.s. Ceylon had j ustleft for M inikoi owing to a wreck on that atol l. This necessit ated a delay of seven weeks ,before we could hope t o sail thi t her , a deten t ion further increased to eight weeks owingto stress of w eather. Mr Borradaile accordingly proceeded to the J afl

‘

na Peninsula,where

he spent a mon t h in familiarising himself wit h the li fe and conditions on coral reefs.I mean t ime prepared our st ores , and made arrangements for t he Maldive cruise

,subse

quen t ly v isiting the raised limest one hi l ls and area of the north of Ceylon‘. After returningt o Colombo I traversed the entire coast be t ween Negombo and Dondra Head , a distanceof abou t 120 miles. Mr Borrada i le j oined me, and we then settled down for a fortnightat Weligama—the Religam of Prof. Haeckel—where there is a deep bay wit h reefs ofsmal l size across its ent rance. The rich variety of animal life on t hereefs bot h hereand off the Jaffna coast as compared t o t he reefs of the Maldives and M inikoi is a mostnot iceable feature.

Aft er a tedious week’s detent ion in Colombo , we final ly left Ceylon for M inikoi onJune 17 t h , experiencing a very heavy north - west gale the whole way ; in spit e of thisthe sea one night was whit e with phosphorescence

,a very unusual phenomenon in these

wat ers. We located oursel ves, and built a bungalow at a dist ance of about one - third ofa mile from t he south - west end of M in ikoi island , under the shadow of the light house, t heboat belonging to which was through t he kindness of Capt . Cbanner , R.N.

,freely placed

at our di sposal. The island here is about 470 yards across between tide - marks,and

a broad ride has been cut , giving the only open space of any size in t he island. Thevegetat ion is ex t remely dense

,and forms a low jungle of Pandanus, H ibiscus, Hernandea ,

Ri cinus , coconut and ot her t rees , wit h Pemphi s acidula,Scaevola koenigz

'

i and Tour -nefor teaa rgen tea. on the shores. At t he sou t h - west end of the island is a shrine

,the grave of a

holy,Moslem sheikh , connect ed by a good , shaded path with the vil lage in the centre of

1 VideRepor t B r i t.Asa ,pp.400—2, 1900.

2 J. STANLEY GARDINER.the island. A small set t lement formerly exist ed round the shrine

, but i t has long beenabandoned

,and the j ungle near it is now far t hicker and less t rodden than elsewhere.

The land has t he same charact er up to t he vi l lage, bu t furt her nor th i t i s much moreopen

,and can indeed be t raversed almost anywhere‘. Our situation then was not unnatural ly

the best possible for t he land fauna, on which t he ligh t house lamp t oo had doubt less noinconsiderable influence. The open ride formed bot h by nigh t and day our best col lect ingground

,sugaring never mee t ing with any measure of success elsewhere.

For marine observations and collecting ,we had within a stone’s t hrow of t he house

on t he seaward face of the island a broad reef- fia t , on which t he sea continual ly breaks.Towards t he nort h t his gradually narrows, but westwards broadens, and continues round t he

at ol l. A broad boulder zone, which can be waded convenien t ly up t o half- t ide,extends the

main (Mini koi ) is land inside the reef- fia t to W iringi li and thence to B agandi and round t he

atol l. These isle t s are mere rocky patches , t he former with a few coconut trees , under theshade of which strangers are buried. Towards t he lagoon t here is a great sand - fiat , exposedat spring tides from 100 to 200 yards from t he beach. The situat ion was also chosen

,as

during t he summer mont hs the sou t h - wes t monsoon blows , t he effect of which I wishedparticularly t o s t udy. Unfortunately t he monsoon of 1899 was very abnormal, t he prevai lingwinds coming from west to west - nort h - west unti l the second week in August

,when the

proper monsoon commenced , bringing heavy rain in i t s t rain. The lat t er made workex t remely difficult and unpleasant ; t he bott om could nowhere be seen on account of t he

surface disturbance ; bott om living animals con t racted , or ret ired into the sand or o t hershelters ; t he surface fauna sank to considerable dept hs.

The disadvantages of the posi t ion lay in the considerable distance of the house fromthe vi llage and from the nort h passage in t o t he lagoon , through which alone access to theopen sea could be ob t ained in t his monsoon. Natives had to be hired from the vi llage foreach several job

,and i t was too far for the children or fishermen to bring any s t range

animals they might find. The wind being dead in our teeth , and t he numerous shoalsmaking short tacks necessary

,i t was difficu lt to visi t t he nort herly reefs of t he atol l

,and

on no occasion was I enabled to approach t hem from seaward within about 200 yards.

During t he months of July and August a heavy eas terly swel l came up wit h largerol lers

,t hree t imes dying down and again regaining force. This swe l l was very abnormal

a t the t ime of year, and apparent ly was due t o some cause completely outside t he ordinarywinds and current s. On enquiry I ascertained t hat it was also observed on the east coast sof Ceylon and India and on t he large Ocean Liners proceeding from Ceylon to Albany.Subsequently I found t hat i t had been fel t throughou t the whole of the Maldives ; in Suvadivaand Addu i t did considerable damage

,sweeping over islets and land , which had never been

1 Owi ng t o an old arrangement theproduceof ha lf theis land i s deemed to belong to theBebeof Cannanore. An

arbi t rary l ineof divis ion ex is t s near thevi l lagewi th gates

and guards, w ho rigorously ex act three- fifths of thecoconu t sgathered south of thefence. In th is port ion no t imber of

largesi zeor old growth ex is t s , thewholesurfacea t onet imehaving been cleared and planted. Subsequen t ly on theholdof theBebebecoming weakened or rela xed

,veget a t ion w as

a l lowed aga in to assumei t s sway, resul t ing in thepresen tdense, j ungly growt h. On themanagement of theBebe’sdomin ions being undertaken by theB rit ish Government , the

old l ineof divis ion as found w as reta ined permanent ly withmuch inj us t iceto the inhabitan t s , as i t had been formerlyper iodical ly open to revis ion. S ince that t ime the north

half of theis land has been very rapidly wash ing away, wh ilethesouth hal f has

,if any thi ng, been increas ing somewhat

in breadth. Thenort h half cannot now annua lly supportone- th ird of the presen t popula t ion , wh ile t he south has

becomea dense j ungle, rapidly going t o was te. I t producesannua lly under thepresent sys temonly a few hundred rupees’wor th of coconut s, wh ich thegovernment might well commutefor a fixed annua l charge.

INTRODUCTION. 3

affected before. The origin of t he swel l can only, I consider, have been due to submarinevolcanic dist urbances probably towards the Eas t - Indian region.

During t he first five weeks of our stay a t M inikoi, while I was engaged in a survey

of t he land and shores,Mr Borradaile occupied himself mainly wi t h a t horough study of the

land Crust acea‘. An incau t ious exposure on the reef, whi le collect ing, t hen laid him up withsunstroke

,so t hat I had no opt ion but to send him t o Ceylon , whence he was ordered t o

return as soon as possible t o England. For t he remaining eight weeks I w as absolut elyalone

,being deprived by il lness even of my Singhalese servant s.

212 21

6?V er y S h o a l o

2}

O

n u m er o u s c o r a l. h ea d .

29 x146 so 22

F IG. 1. M in ikoi At oll , from theAdmiral ty Char t. (W i r ingi li , na t ivenamefor Sma ll Fox

I ret u rned t o Ceylon in the middle of September,Mr Forster Cooper shortly aft erw ards

j oining me from England. After some unavoidable delay we left Colombo on Oct. 18 t h forMale

,t he capi t al of t he Maldives and residence of the Su ltan. Owing to a succession of

accidents we did not arrive unt i l Oc t . 23rd,w hen we at once landed our stores

,transferring

sufii cien t for a t hree months’ cruise to our schooner. The latter was lent t o us by H is

H ighness t he Sult an ; she was a vessel of about 16 t ons, built in the islands, of coconu t

1 VideM r Borradai le’s accoun t in thesamepar t of th is publica t ion.

1mandi fulu

Makunudu .

M i laduma

Maihlos

madulu

Goi furfehendu

Male

N Ni l

MulakuS.Nilandu

Kolumadulu

Hadduma t i Ghang

na Mulaku

Frc.2. Ma ldiveGroup, showing the

route of the ex pedi t ion. Sca le60mi les to 1 i nch.

J. STANLEY GARDINER.

wood , moderately seawort hy, but not laying wit hin six pointsof t he wind. The Sult an also appoint ed Hassan Didi Velanamanikofanu , his third vizier, t o accompany us, and gave orderst hat every faci lity should be granted to us. Aft er presen t ingour offerings t o t heSul tan and his viziers, we sai led from Malefor G oifurfehendu (Horsburgh) atol l.

We at once had a housebuilt on Goidu island , from which ascent re wevisited al l the other land of the at ol l and the great erpart of t he reef. A stay of al t oge t her eleven days was made

,

and everyt hing was unpacked and properly st owed on theschooner ; t he dredges and instruments were overhauled , andindeed all preparations werecompleted for ' thework in Mahlos

and other at ol ls. The reef- animals were col lect ed and preserved,

being sent by na t ive boat to Male to await our arrival. On

leaving M inikoi I brough t wit h me t w o boys , whom I hadt aught t o collect in that atoll ; these I largely employed inGoidu and subsequen t ly in other islands in collecting t he landfauna and flora.

The remainder of Oct ober, 1899, was spen t in S.Mahlos

madulu. This group of reefs real ly consists of three at ol ls,a

smal l central one separat ed from larger on each side by narrowchannels of over 100 fathoms in depth. The three lie on ashallow bank

,which tapers t o the north but has a broad base

to the sou t h - eas t. The plateau is studded all over wit h reefs,

t he outside ones forming a chain round the '

perimete-rs of t het hree parts. The reefs along the west side of the bank are forthe most part ring - shaped , smal l at ol ls (a t ollons or faro) withdeeper water (the lagoon or velu) in t he centre. On the eastand sout h sides , however, isolat ed islands with fringing reefsmost ly form t he boundaries. The general dept h of t he atol lsis abou t 27 fat homs, most of t hechannels between the numerousencircling reefs having over 20 fat homs.

The weat her during our stay in S. Mahlos was ex t remelycalm

,our vessel indeed being t owed by boat s from island t o

island. This was singularly unfort unate , as usual ly st rong windsmay be depended upon in November. We had hoped to systema t ically dredge a large number of the deep channels betweent hereefs, t hat edge the at ol l. As t heir general dept h is about25 fathoms, t his was found to be impracticable, rowing boat snot having suffi cient weigh t t o carry even the smal lest dredgesalong a rough bot t om a t this dept h. Accordingly we confinedourselves t o a traverse of t he whole south of t he at oll. Our

first anchorage was off Turadu , an island si t uat ed on the rimof a somewhat i l l - defined faro a t the sout h - w es t corner of t hebank. We Visit ed every part of i t s reef and col lected a fewanimals. The lagoon (velu ) of thefaro was dredged , yielding

INTRODUCTION. 5

Asymmetron and P tychodera from 20 fathoms. The island itsel f proved most interes t ing. I t s

rocky barrier of beach - sandst one had in 1896 been overlapped by t hewaves of a cyclone.These att acked t he sand behind , eat ing deeply int o the island , with the resul t t hat t he

beach - rock has been left in lines many yards from t he shore. The nat ives have now erectedbreakwaters round a great part of the island—and also a new mosque—but in spit e of t heseno trace of i t is likely to be left in 20 or 30 years’ time , unless some considerable changein the currents or reefs alt ers its conditions.

From Turadu we visited al l t he reefs t o Mabaru, the most east erly point of t hew hole

group,anchoring at Hi t adu,

Heddufuri , Mahrus and Duravandu. At Cumfinadu we foundsome large rocks , s t anding up in the lagoon wel l inside t he boulder zone ; B onelli a wasl iving on t he reefflat , and Ptychodera was the most abundant form of life on the shoresof t he island.

We final ly left for N.Mahlos on Nov. 29t h, but , meeting wi t h strong currents to thewest - sou th - west

,we only fetched Kuderah - Heclu in the central at ol l that evening. However

we reached Fa inu in N. Mahlos on t he fol low ing day , and remained t here at anchor for t hreedays

,which were devot ed to dredging and an examination of t he islands of Fainu

,Kenurus

and Ingurahdu ,and B erriam - furi faro. IVe then separated, Mr Forster Cooper dredging with

the schooner along the east side,and examining its islands and reefs. I meant ime embarked

in a small open boat for the wes t ern side , where the lagoon of the atoll is fi l led up witha perfect maze of smal l reefs and shoals. I first visi t ed five of the lagoon islands

,and

then,a strong north - east breeze setting in

,worked up along t he edge of the at ol l

,sai ling

from dawn to dusk,and anchoring a t night to leeward of the nearest reef. I examined

all t he reefs and islands, and sounded t he velu (lagoons) of al l the faro (a tollons) along t he

western rim,rejoining the schooner on Dec. 11 at the north of the atol l. Unhappily a series

of col lections,made by mysel f and my boys

,to i l lustrate the populating of sand - banks by

both animals and plant s, was ruined by my capsizing our fishing boat near Cunderudu.

On Dec. 12t h, aft er taking in wood and water, we sai led for M iladumadulu,a similar

bank t o Mahlosmadulu bu t with relatively far fewer reefs and a less determinate rim.We anchored the same nigh t a t G ut hardu, and at daybreak made sail towards t he eastside of the at ol l. Ow ing to a s t rong sou t h - westerly set of the current we took t w o daysin reaching Dureadu in the middle of t he bank

,a dist ance of seven mi les: This is the

island of a round faro,one mile in diameter ; i t s lagoon has 19 fat homs of water , a dept h

which makes t he fare peculiar among al l i t s fel lows in t he Maldives. We landed t he

same night for firewood and wat er,but , none of the latt er being obtainable, w ere compel led

to sai l at dawn. A strong north - east gale coming out,we stood up t he atol l and wat ered a t

Rymaggu ,anchoring that night at Furnardu

,a large island on the eas t edge of t he at ol l.

The is lands of t his rim of M iladumadulu tend to be closely fringed by t he reef on allsides

,and to have a kuli (shal low lake , French ba rachoi s) in the cent re, surrounded by

mangroves,through which the sea has in some found access. While Mr Forst er Cooper

dredged wit h the schooner down t o Kendikolu,I visit ed ten of t hese islands in a fishing

boat ; two were very smal l , t hree had definite kuli or else mangrove swamps (Ekasdu,a

large lake swarming with a species of Leander ), and four were crescentic in shape, t heirkul i now open to t he sea. Kendikolu is one of t he largest islands in t heMaldives, beingtwo - and - a - half mi les long by two - t hirds of a mile in breadt h. There are four kuli downthe centre surrounded by mangroves

,which abound in rai ls ; on t heir surfaces we saw

6 J. STANLEY GARDINER.a few duck

,while t hei r waters , which are quite fresh , four to five feet deep

,teem wit h

smal l fish. However , as Ramazan , t heMahommedan fast mon t h , was approaching, we had tohurry on , and only stayed two nigh t s. We accordingly , on Dec. 19t h, dredged down toLandu , obt aining a large variety of sponges and Polyzoa wit h a quanti ty of red Polytremaand some nul lipores , in addi t ion to corals , of which the black Dendrophylli a fr amea was veryabundan t in 12 to 20 fathoms off the reefs. We vis ited and dredged Ma and Eddu farowit h t hree t o four fat homs of wa t er in their velu , obtaining a few Cephalochorda, and examinedsome of t he cen t ral is lands of t heatol l. At Manadu we caugh t a single specimen of Typhlop s ,which is eviden t ly very rare in t he archipelago , since i t has no native name.

Fadififolu at ol l was reached on Dec. 23rd ; i t differs from t he preceding in being a t rueatoll

,having a well - defined, encircling reef, especially to the east , and an open lagoon w i t h

but few shoals. We first moved down the eas t side,examining the reefs

,but t hen dredged

across t o Innaw ari. The natives, alt hough not act ively host i le,w ere very unfriendly, so t hat

on Christmas morning we sai led down t o Na ifaro. The w inter rains now commenced, andcontinued in t ermi t tently with heavy squal ls from the nort h - east for a mont h

,making

navigat ion among reefs diflicult. The schooner too was decidedly unpleasan t,as t he whole

of our cabin accommodation had to be utilised for storing our books,instrumen t s and

various col lect ions. We remained at Naifaro four days,rest ow ing and repairing our vessel,

as she had been somewhat st rained in t he recent heavy weather. The time was occupiedin a survey of the neighbouring islands and reefs, and in thoroughly dredging some of t hepassages into t heat oll.

The shores of al l the islands a t the nor th end of Fadiffolu I found abundantly st rewnwit h t he shells of Sp i rula. Enquiring of t he natives as t o its “ fi sh

,

” I was surprised to

receive an accurate description of it. I t appeared t hat the animals were ext remely abundan tin January of 1897 in t he channel t owards M iladumadulu. “ They float on t he surface

,

and may be pick ed up wit h t he boat - bai ler. They are never seen inside t he atol l,but

periodically occur in t he nort h - east monsoon in t he open sea. None were seen in 1898.

I offered a reward of 50 rupees for t he fi rst specimen , but , alt hough eigh t or nine boat swent out daily during my st ay , I did not secure one. That t he people of Na i faro andInnaw ari really know t he animal is undoub t ed. Y et t his is peculiar, as I did not findon close enquiry t hat it was known to t he nat ives of any other part of t he wholearchipelago, t hough i t s shel l occurs sparingly everywhere. The native name for the shell isma rkana ta lndandi , t he heron

’s key.

Leaving Naifaro we again dredged across t he atoll on a more sou t herly course, subsequent ly cruising along t he eastern side, dredging and examining the islands and reefs.On Jan. 2nd,

1900, we set sai l for Male at oll , anchoring off Helengeli t he sameafternoon.This atol l is in termediat e in i t s characters bet ween Mahlos and Fadi tfolu ; except a t t he

sou t h end i t has pract i cally no islands in t he lagoon. There are evident ly great changesin i t s t opography since t he original survey, t w o islands of t he east ern edge a t least , rest ingon t heir own reefs , having disappeared. There seemed t o be st i l l great er al t erat ions inrespect t o t he shoals in t he lagoon , but in our somewhat dependen t posit ion any realsurvey was impossible owing to const ant int erference from Male. We, however, saw someof t he western reefs on our way t o G oifurfehendu

,and on our ret urn j ourney t o Male ,

which we reached on Jan. 5 t h ,1900,

we dredged down as close as possible t o t he east ernreefs. Further Mr Forster Cooper in t he middle of February made a most successful

INTRODUCTION. 7

dredging cruise of eight days in the atol l, taking 34 hauls, to ascertain t he charact er of t hebottom in every posi t ion

,and I at the same t ime visited most of t he islands and reefs

w it hin five miles of Male.

Ramazan had now commenced , and , dependent as we were en t irely on Mahommedanboys

,i t would have been useless to cont inue our cruise. Our vessel too was in want of

a thorough overhaul,the rigging being very bad. Aft er a couple of days in Male

,we

had a house erected on Hulule , the island of a neighbouring faro, and at once transferredto it sufficient st ores for a stay of four or five weeks’ duration.

Hulu le island is about 1} miles long by 800 yards broad ; i t is about two milesdist ant from Male. The greater part of its surface is covered with coconu t trees

, bu t

a large patch along t he western side has been al lowed to revert t o j ungle. The principalt rees are the banyan , candle—nut and

"

Ca lophyllum, the branches of which abound infrugivorous bats. On account of t he island 's proximity to Male , where al l foreign vesselsfor the group have t o enter, many plant s have been introduced. Few of the frui t s thrive

,

but half - a - dozen bri l liant flowers relieve the everlast ing green. Sweet - smel ling plants,

j essamine,frangipanni

,roses and various herbs , make the proximi ty of t he mosque and Vi l lage

pleasantly fragrant. The wes tern shores are fringed wi th Pemphi s a cidula,t he whit e

,

perfumed flowers of which prove a great at tract ion t o insect s. As some of our boysdelight ed in this work , the land fauna and flora was exhaustively collected. As comparedwith M inikoi , we found the insect and spider faunas to be decidedly poor excep t inbutterfl ies

,although it may be deemed to be thoroughly representat ive of any rich island

in the cen t re of t he Maldives. A cert ain number of insects must have been introduced withthe plants

, bu t the successfu l acclimatisat ion of any considerable number of t he lat ter onlydates back to the erupt ion of Krakatoa in 18 83. Before this time pumice (feng - bo-

ga ,t he

wat er - swimming stone) was not known in the group. Its fertilising propert ies have now beendiscovered

,and in many islands baskets of i t are collect ed and strewn over the garden land.

The capture of two specimens of snakes on pandanus trees at Hulule was of int erest.

The whole faro is 4Ami les long by 11} broad ; i t s lagoonlet , or velu , in the cen t rehas a dept h of 6 to 7 fat homs. The reef everywhere is awash at low t ide

,and

,though

differing great ly in its characters, quite wel l defined on t he side t owards t he lagoon of t helarge atoll. Besides numerous rocks t here are two islands respectively a t the nort h andsou t h ends , Farukolufuri and Hulu le. During our stay every part of the faro was surveyed ,and the fauna carefully col lected for comparison wit h that of M inikoi atoll. Asymmetronwas very abundant in certain places in and around t he velu and a t least t hree species ofEnt eropneust s of two or three genera were found, each in I t s ow n charact eristic environment. Tw o specimens of a remarkable Tha lassema

,2 feet long when alive, were secured

from t he boulder zone , besides a number of specimens of smal ler species. The sand wasremarkably rich in Actinians of many species

,al l of which w i t hdrew in t o the shel t er of

the sand at every rain—squall. Mollusca were not numerous,but Cryp top lax , Chi ton,

andeach of the three families of the Zygobranchiata were represent ed ; autotomy of the footseems t o be a widespread phenomenon , as several forms with the foot thus cut off wereobtained.

We ret urned to Male in the middle of February. Mr Forster Cooper at once sailedfor his dredging cruise in the atol l

,while I remained to carry on a series of observations

8 J. STANLEY GARDINER.

on t he currents in the channels on each side of t he island. During my whole stay Idispensed medical aid freely to the people. At t his time there was in one dist rict of Malean epidemic of malaria, wi t h which I was naturally much brought into contact . Mr Forst erCooper returned t o find t hat I had caught the infect ion. I saw

,however

,our col lect ions

to dat e properly packed for England , and the schooner victualled and equipped. As complications set in and I was gett ing worse instead of bet t er, Mr Forster Cooper on Feb. 28 t ht ook advant age of the chance visit of a Brit ish India Steamship Co ’

s st eamer t o place mein charge of Capt. Pigott, for Colombo. He himself determined to carry out hispart of t heproj ect ed work, and sailed from Male on March 3rd for the sou thern atol ls.

I cannot speak too highly of t hepluck , determinat ion and resource, shown by Mr Fors t erCooper, who had had no previ ous ex perience of t he t ropics. He worked his nat ive crewin a manner

,which I have never seen excel led even in the Pacific

,and took no less t han

8 8 dredgings in five different atol ls. However, I append Mr Forster Cooper’s report

,which

may be allowed to speak for i t self

On March 3rd,accompanied by Mahommed Didi as interpreter and represent ative of

t he Sult an, I crossed over t o S. Male, in which t hree days were Spent . G urahdu island andreef were visited, but neither land nor reef in t he at oll appeared to e x hibit any novelfeat ures. The group consisting of a series of almost isolated reefs, I dredged principal ly inthe outer passages and in the cent re to ascertain t he differences in t he bottom—fauna.The hauls in t he centre were very unproduct ive, bu t in the channels a large quantity ofthe same sessi le forms, as in t he nort hern atol ls

,was brought up. Just inside t he northern

passage a coral , Goniopora stokesi,was obtained ; i t forms round heads on t hick stalks,

covered by an epit heca,which is completely buried in the sand

,or mud.

On March 6th we sai led on to Felidu a t oll,where we vis ited Alima t a and Tinadu

islands alone,at nigh t anchoring generally to the nearest reef. We remained eight days

,

bu t on ly t ook 18 hauls of t he dredges owing to head w inds,strong curren t s across t he

atol l and dead calms. Indeed throughout t he whole cruise the unfavourable winds and t he

poor sai ling quali t ies of the schooner prevented us from surveying as much of the atollsand from t aking as many dredgings , as I would have liked. We at t empted for t hreesuccessive days t o get up the almost complet ely enclosed eastern horn of t he at oll

, but

were unable t o t ack up more t han half- w ay. The lagoon in it is open wi t h few reefs orShoals, and has a general dept h of 40 fat homs ; i t s bot t om is hard , covered with sand , andabsolutely unproduct ive to the dredge. In the rest of t he at oll weed and broken shel lswere found t owards t he centre of t he lagoon and rubble in the channels. Dia ser is is verycommon almost everywhere and i t s skeleton forms one of t he chief constit uents of t he rubble.

“ We fet ched Mulaku at ol l on March 14th,and remained six days

,taking 17 dredgings.

The at ol l is remarkable for the broad , almost con t inuous reef along its eastern side. Therea re a number of very narrow , shallow passages t o t he nort h

, but from Maduveri t o Curai lle

a dist ance of 30 miles, the reef is con t inuous save for a single passage opposit e Mulakuisland. All t he islands lie on the seaward side of t his reef ; ext ending down i t s cent re isa series of long

,linear velu (small lagoons), which ofi" Ra imandu is double , t w o series lying

parallel t o one anot her and the edges of t he reef. In dredging I ran t w o lines across t helagoon

,and t hen moved down along t he east ern side. Among ot her forms we collect ed a

l I cannot sufficient ly ex press my indebtedness to th is gent leman for all h is kindness t o me.

https://www.forgottenbooks.com/join

10 J. STANLEY GARDINER.I t has a wel l - defined rim wit h passages a t int ervals

,and is about 34 miles long by a

li t tle less in bread t h ; i t s lagoon has a maximum dept h of 50 fat homs. AS Addu at ollwas t o be our t urning point , we now spen t only two days in Suvadiva ,

dredging andsounding along t he east side of t he lagoon. We anchored with banked fires for two nigh t sa t Ni landu and G addu

,and I further

,leaving the dredging t o my companion

,visi ted

W iligi li and Konda i.

Addu is a small atoll,10 miles long by 6 broad

,lying about la t . 0° 40’ S. I t s reef

is perfect except for two smal l passages t o t he north and t w o larger ones t o the sout h.The lagoon has a max imum dept h of 36 fathoms ; i t is fairly open in t he cen t re , bu tagainst t he encircling reefs has a. perfect maze of coral heads

,arising from 7 t o 10 fat homs.

I t is noticeable that t he greater part of t he circumference of t he atol l is surrounded byland. The “ Equatorial Channel ,

”

48 miles across,separat es Addu and Suvadiva. In its

centre is t he island of Fua Mulaku, said to have a deep kuli ( lake) in its midst ; it istwo miles long by one “broad

,and has a fringing reef on ly. On our passage t o Addu , and

subsequen t ly on our return j ourney ,we tried t o visi t it, but t he heavy sea made both

anchoring and landing quite impossible. We remained a t Addu unt i l April 15 , examiningt he reefs and islands

,in both of which there have been great changes since Moresby’s

survey. We also checked t he soundi ngs on t he char t , and t ook 14 hauls of t he dredgein t he lagoon and on the outer slopes. In one of t he lat ter from 40 fat homs we obtaineda large quan t it y of Heliop ora coerulea with almost colourless corallum. The veget at ion oft he islands was far more luxuriant t han any we saw in other parts of the Maldives , butthe land fauna was very scanty as compared wi t h Hulu le. The animal life of t heencirclingreefs seemed to be both poorer and less varied t han to the north ; t he growth of fixedforms of life in t he lagoon was , however, decidedly lavish.

Revisit ing Suvadiva on our return j ourney, we dredged and sounded within the lagoonalong t he sou t h

'

and west S ides , anchoring for t hree night s a t G addu,Nadalle and Havara

Tinadu,beyond which I did not land anyw here. On Apri l 19th we entered Kolumadulu,

having in t hemorning coasted along t he east side of Hadduma t i. W e dredged the passage,as we ent ered

,and dropped anchor at Buruni , a t once going on shore.

On April 20 we ran a line of soundings across t o S. Nilandu, an atoll somewha tsimilar t o Male. We twice t raversed t he lagoon , anchoring that night at Rimbudu. Sixdredgings were t aken in 19 to 35 fat homs ; t hey proved to be in t heir resu lt the richestseries t hat were obt ained in the whole archipelago, al t hough I do not t hink we foundany forms of which we had not previously preserved specimens.

On Apri l 21 we sounded t he channel across t o N. Nilandu atol l, which we crossed.We then ran a line of soundings t o W a t t aru at oll across t he cent ral basin , which hasvery general ly been supposed t o owe i t s origin to the lagoon of a much larger at ol l,now complet ely los t . W e furt her sounded t he channel between Mulaku and W at t aru at olls,anchoring for the nigh t off Rak idu in Felida at oll. On Apri l 22nd we sounded thechannel between Felidu and W a t t aru at olls

,and then ran a second line across to t he

w est ern chain,abou t 1 5 miles nort h of t he first. W e sounded t he channe l between N.

Ni landu and Ari at olls,con t inuing northwards sounding a t in t ervals t o Mahiaddu in t he

lat t er atol l, where we dropped anchor. On April 23rd we made a st raight course for Male ,putting down our t h ird line of soundings across t he cen t ral basin.We ran our line out

INTRODUCTION. 1 1

along t he channel between N. and S. Male atol ls , ob t aining finally a depth of 1005 fat homs ,two miles S W . i W . of the S. point of t he reef of H ulule faro. That nigh t we anchoredoff Male and, aft er taking our col lections on board , bade good - bye to t he Maldives onApri l 25 , 1900.

The col lect ions may be al lowed t o speak for t hemselves in t he subsequent part s ofthis publication. A word is necessary as t o the dredgings

,of which 273 were recorded in

the Maldives. They were in t ended to ascerta in the character of the bo t tom wit hin theatolls as well as i t s fauna in every position , in w hich t he physical condi t ions might vary.Natura l ly a considerable number were under these circumst ances absolu t ely barren , but al lserved t heir purpose. Seven dredges were used , both rec t angular and triangular, of Naplesor Plymou t h models

,as wel l as an ot t er and three beam trawls of 3 , 4; and 6 feet.

The lat t er were made t o my own design,an adaptat ion of Prof. Agassiz’ model ; they were

for t rawling from the st eamer weighted w it h fire—bars ; their nets were of coconu t fibre

(sinnet or coir) , and were made by my boys. Swabs of hempen rope , three feet long, werefound best adapted to our work ; they were always thrown overboard, w hen we were atanchor

,about one fat hom of rope being al lowed beyond t he depth , i f t he bo t t om was

fairly level. For sounding from the schooner and small boats,besides regular lead lines ,

we used loosely spun cod - fishing l ine,as recommended by Mr J. Y. B uchanan. On t he

steamer we had an old Lucas deep - sea machine , which was lent us by t he Admiral ty ,Mr Lucas kindly providing abou t 3000 fathoms of w ire. Of leads we employed t he Telegraphic Const ruction and Main t enance Company’s snapper

,and valved leads of theAdmiral ty

pat tern. Ot her apparat us and ins t rument s wi l l be referred to where necessary in the accountof the work of the expedition.

J. STANLEY GARDINER‘.

[Note The collect ions of plants have been presented to the Royal B otanical Gardens ,Peredeniya , Ceylon. A ful l report on them wi l l be shortly published in the “ Journal ofthe Gardens by Mr J. C.Willis, the Director. J. S. G ]

1 Th is account has been checked by Mr L.A.Borradaileand Mr Fors ter Cooper.

THE MALDIVE AND LACCADIVE GROUPS W ITH NOTES ON OTHER

CORAL FORMATIONS IN THE INDIAN OCEAN.

B Y J . STANLEY GARDINER, M.A.,Fellow of G onvi lleand Ca ius Collee

and B a lfour S tuden t of theUniver si ty of Cambr idge.

(With Plates I. and II. and Figs. 3

CONTENTS.

CHAPTER I. THE CORAL REEFS OF THE INDIAN OCEANTopography—Soundings—I.M ALAGASY S ECTION—Madagascar—Comoro I s.—Europa I.—Prov idenceI .—Cosmoledo I.—Aldabra I.—II. S EY CH ELLE S ECTION—Rodri guez—Reun ion—Na za ret h B ank —Saya de Ma lha Bank—Seychelle I s. -AmiranteIs.—III. CHAGOS S ECTION—IV. MALDIVE SECTION—Channel bet ween Ma ldivesand Chagos—India—Ceylon.

CHAPTER II. METEOROLOG ICAL CONDITIONSGenera l—W inds (Monsoon and H u rri canes)—Current s (Genera l, across At olls, Depth,

Effec t s of Reefs on D irect ion)—Tides—Ra infa l l—Tempera t ure.

CHAPTER III. TH E ATOLL or M INIKOI

SEC. 1. G ENERAL—Ou ter S lope—Topography—S EC. 2. LAND.—Rocky area—Elevat ion—Sea - bea ch—Lagoon - bea ch—Eros ion—Pi l ing up of Sand—Tunda Po intReef—rock—D is t ribu t ion and S t ru ct u re of Conglomera te Rock—Evidence and

Ex tent of Upheava l—Sandy Area—Addi t iona l Ev idenceof Upheava l—W iringili

and B agandi I s.CHAPTER IV. TH E ATOLL OF M INIKOI (Con ti nued)

S EC. 3. TH E ATOLL REEF—Kodi Point (Terra ces and Conglomera teMasses)—MouRambu Point—Ko- Var i B ay—Tevera t u Point (Cora ls , F iss ureZ one, B u t t resses)—Ligh t li ouse—Bet ween W i ringili and B agandi I s. (det a i led Descript ion and

organ i c Life of t heF issure Z one and Fla t )—The Boulder Z one (Descript ion,Conglomera te Masses , organ i c L ife

,et c.)—SEC. 4. TH E LAGOON—Current s ,

Tides and Plank t on—Reef- fla t off B on i - Kodi—Sand- fla t (Descript ion and organ ic

CORAL REEFS OF TH E INDIAN OCEAN; 13

Life in differen t Pos i t ions)—Lagoon Cora l Reefs (D ist ribu t ion, Descript ion,dead Reefs

,et c.)—Deep W a ter—SEC. 5 . TH E H ISTORY or M INIKOI As AN

ATOLL—E leva t ion—A Perfec t A t oll—G a in and Loss—A Cen tury Ago (na t iveEv idence)—Reconst ruc t ion of Land and Reef.

(To becont inued.)

CHAPTER I.

THE CORAL REEFS OF THE INDIAN OCEAN.

THE t opography of the Indian Ocean gives but li t t le aid to the student in at t empt ingto det ermine the character of t he foundations of the various island groups. Far less indeed isthis the case here than in the Pacific Ocean

,where the trend of each of the larger groups

has a t endency to run somewhat paral le l to t hat of others and to continental land , sug

gest ing t hat they lie on lines of weakness of t he earth’s crust , lines o f former volcanicact ivi ty. The t wo oceans are alike in that the west ern part of each is almost bare ofcoral islands

,the Pacific absolutely so

,

“ the Indian Ocean with Cocos - Keelingl and Christmas

islands 2 alone. Of t hese t he former Is the t ype,on which descript ions of most at ol ls have

been mou lded, and the lat t er is an ancient a t ol l , which has undergone great elevation , theonly one indeed in the whole Indian Ocean. This lack of raised reefs is remarkable, ascompared wit h t he Pacific Ocean , and can only be taken as an indication of the absenceof any considerable volcanic activity in modern t imes , at any rate between Madagascar andIndia.

There are only two changes in the past in connection wit h the topography of t he

Indian Ocean , abou t which t here is any general consensus of opinion , i e. a former waterconnection w i t h t he Atlan t ic Ocean ,

and a land connection between Madagascar and Africa.The permanency since the Jurassic period of t heArabian Sea and of the great basin bet weenChagos and Rodriguez on the one hand and Aust ralia on the ot her has not been seriouslyimpeached , so t hat there is only left a belt betw een S. Africa and S. India, and i t is inthis that al l t he considerable coral reefs lief The general dep t h of thew hole Indian Oceanis remarkably uniform

,about 2500f (fat homs) , a single sounding only of over 3000f (3097

t o the S W . of Christmas island ) having been obtained. What however is sti ll more noticeable is , that not a S ingl e sounding N. of t he line between Cape S t Mary (S. Madagascar)and Cape Leenw in (S.W .Australia) has been run of less t han 2000f. ex cept near cont inen talland , or c lose to islands and shal low ree fs , t he existence of which is known a t the presentday. The soundings, such as they are , alt hough not quit e close enough, give s t rong in

dica t ions , t hat the bel t between Madagascar and India is cut by a depth of over 2000fin three places, i .e.

'

bet ween the Maldives and Chagos , bet ween t he lat ter and Saya de

1 Vide“TheS tructureand Di s tr i but ion of Cora l Reefs , 9 Vide “ A Monograph of Chr is tmas Is land, Indianby Chas.Darw in (1889) and TheCocos - Keel ing Is lands " Ocean.” B r it ish Museumby H.P.Guppy, S cot t i sh Geo. M ag.

Ban a l

da Ind ia A!

l8 4 5

m 27m

z¢7 z

Amiranw Bank1 15 0 PPM 4 1

h C "

228 0

F IG.3. Indian Ocean , Western port ion.

CORAL REEFS OF THE INDIAN OCEAN. 1 5

Malha Bank,and between Farquhar atol l and Madagascar. These channels divide the coral

reef areas into four sections , which may be termed respectively t he Malagasy,Seychelles

,

Chagos, and Maldive 1

1. MALAGASY SECTION.Madagascar itself displays a nucleus of granite running from end to end of the island ,

broken only by pat ches of recent ly act ive volcanoes. No raised rock of coralline origin isknown

,bu t nummulitic limestones of the eocene period form a belt along the western side.

The reefs round the island are of inconsiderable extent , most ly mere isolat ed fringing banks.The former connection of t his island wit h Africa is indubitable, and its st ructure indicatesthat i t has experienced several changes of level since t hat connection was broken down.The Comoro islands, a volcanic group NW . of Madagascar, have only fringing reefs exceptMayot t a ,

w hich possesses a barrier , 3 to 6 miles distant , wit h upwards of 35 f of water inits lagoon. Two smal l high volcanic islands arise in the barrier itse lf, indicating that i t sformat ion was probably sui generi s, on a bank formed of volcanic debris from the cent ralisland.

Of i solated coral format ions there are to the south about lat. 22° in the centre ofthe Mozambique channel Europa island

,50 to 80 feet high

,with low cliffs and sandy hom

mocks,and a smal l fringing reef, and the Bassas da India bank , a round atol l—reef 9 miles

in diameter,complet ely enclosing a central lagoon of shal low dept h. B etween the Comoro

and Cape Amber (N. Madagascar) are a series of banks, of which the G eyser is to al lint ents an atol l with a wide passage t o the S E ; G lorioso has two islands, for which the

descriptions do not indicate any elevation.

To the north of Madagascar , about lat. 9° are a series of banks and islands

,

which Show an unmis t akeable and very uni form elevation of something over 30 feet. Of

these two banks to t he east Providence and Wizard never probably had an atoll form,

they have now merely sand cays Providence two smal l is lands as wel l) and isolated rocks ,w hile St Pi erre in their vicini t y is a raised island , one mi le in diame t er ; i t s elevationwou ld seem to have been but very recent

,as it is described as having no fringing reef

,

the sea directly breaking on its coast. In addition there are three indubit ably raisedat olls , Cosmoledo , Farquhar and Aldabra. They differ from most S imilarly elevated atol ls

,

in that their rock sti l l ret ains even a t t he surface its organic structure. The lagoons oft he original atol ls can in each be traced

,and t he greater part of their basins appears to

be covered with water at high t ide, although t heir mud - banks increase in extent in the

order named. I t is remarkable that t he lagoons of al l have maintained , or obtained passages wit h a few fathoms of w ater to the sea

,and t hey would appear t o be regaining

t heir former ex tent . Aldabra, 19 miles long by 7 broad, has save for two smal l passagesa perfect rim of land, covered by jagged coral , while i t s lagoon is surrounded by mangroves ,which are said to be everywhere “ eat ing the coral away

1 The sou thern is lands of the Indian Ocean (Prince them by over 2000f. They must be regarded rather a s

Edward, Crozet , Heard, Kerguelen , Ams terdam and S t Paul) ou t l iers of thegreat Antarc t ic Con t inent , s incethey all res t

areseparated by at least 1200miles ofocean from thenearest on thesomewhat shal lower bank , wh ich ex tends nor thwardsislands of cora l forma t ion , or with cora l reefs of any sor t. from i t s shores.They are all volcan i c, and can haveno conceivable con.

2 “ Is lands in theSou thern Indian Ocean. Admira lty ,

nec t ion with theis lands to thenorth , being separa ted from p.366, 1891 , from wh i ch also much of the informa t ion in


2. SEY CHELLE SECTION.This section comprises a semicircle of reefs and islands, ex t ending from Reunion to

t he Seychelles and t hen southwards t o the Amirantes and the Alphonse - Francois islands.Rodriguez is an ou t lying basaltic island to t he east of Maurit ius , wi t h a lit t le raised coralrock ; t he reef is fringing or sligh t ly barrier wit h low coral i slet s l. Agalegas and Tromelinbe t ween the line and Madagascar seem to be of coral format ion with fringing reefs ; theyare probably sligh t ly raised , t hough no definit e evidence on t his point i s available. Thesethree islands may wel l al l be separat ed from the rest by over 2000f. of water, but the

main line so far as t he soundings go would rat her appear to be on a bank wi t h a t

most 1000 to 1 500f. of water.Reunion to t he sou t h is a volcan i c Island , feet high , with fringing reefs in

places,whi le Mauritius

,also volcanic

,2700 feet high , has bot h barrier and fringing reefs,

which were visit ed by Darwin. Tracing t he l ine t o the nort h Cargados Carajos is acrescent—Shaped reef

,awash

, 30 miles long, in one place 4121 miles wide

,wit h 7 low

,coral

islands,which appear t o be washing away ; it has a shal low bank e x t ending 16 miles off

t o t he west,but has in no way begun t o assume an atoll form. The bank known as

Nazare t h probably ex tends down wit h.shal low water (under 50f.) and includes the last

named , giving a tot al lengt h to t he bank of 280 miles ; t he Shallowest sounding is 14 f.,but the whole is i l l - surveyed ; t here does not appear t o be any deeper cent ral part . The

Saya de Malha,1 50 miles further nort h , consist s real ly of two considerab le banks which

are probably not separat ed by more than a few hundred fathoms from the last , as comparat i vely Shallow soundings extend out from each for a long distance. The northern bank,75 miles long by 23 broad , has a ring- shaped reef, abou t 12 j. deep (shoalest water 5 f)with a cent ral basin of 40f., while t he sou t hern is 140 miles long by 120 broad, wi t ha rim t o t he N. and E. at 12 f but only a little Shoaler (25 f.) t o t he S. and W .

,

and a cent ral basin abou t 50f. deep , a great high t in one place running in from t he S.with 60 to 70f. The t w o banks are remarkable for the great breadt h of t heir rims,about 7 miles in t he N. bank and 12 miles in the sou t hern. A further smal l bank t o

t he S E. has one sounding of 5 7 f. in t hecent re and rim 25 f. deep.The Seychel les lie about 200 miles N.W . of t he last , with no bot tom soundings in

the channel between. They consist of ten high grani t ic islands, most ly wit h fringing reefs,on a bank 35 f deep , 200 miles long by 120 broad. The northern ou t l iers, B ird and Dennisth is sect ion w as obt a ined. Most of thesurveys of isolatedis lands in theregions compr ised in t h is and thenex t sect ionappear to havebeen madeby Admi ra l S irW .W harton

,

a most ex per t and careful observer. For informa t ion on

Madagascar consult “ M adaga scar , Maur it ius and theother

Eas t - Afr ican Is lands by Prof.Dr C.Kel ler, 1901 , w ho givesa full l is t of references.

1 Vide The Phys i cal Features of Rodriguez ,” by

I s.B ay ley Ba lfour, Phi l.Trans.R.S ., vol.168 , pp. 289292, 1879. The raised coral - rock i s described as forminglarge, cora l l ine, l imes t onepla ins of incons iderableelevat ion.Prof. B ay ley Ba lfour proceeds -

“ Theex is tenceof t hese

masses of corall ine l imes tone indicates clearly a formerlower level of theisland

, and theevidenceof ra ised beachesconfirms th i s. But a considerat ion of thecora l - reefs point s

as clear ly to a t imewhen theis land stood a t a h igher level.The present cora l reef fringes the coast , ex tending about3 mi les on the south - west s ide, but coming close inshoreon theeast . An older reef , however , ex ist s , now quitesubmerged in some places to a depth of over 90 fathoms.Upon i t the present reef rest s, and i t ex tends westwardsnear ly 15 mi les from t hepresent coast , W h ile to theeast

i t s t ret ches abou t 6 mi les. Wehave th us proofs of grea t

and in termi t ten t‘

osc illa t ions of the level of the is land.”I may bepermit ted to point ou t

,tha t i t i s a ssumed t hat

th is submerged reef i s of cora l format ion. I t s contourcorresponds, so far as i t i s known,

nei ther with ex isten tfr inging, nor barrier reefs. I t s or igin i s probably rather to

bet raced to theformer erosiveact ion of thesea on a volcan iccoast.



26 miles long by 16 broad , rim varying with 6 to 12f of water, very imperfect to t hesouth,inside a Shal low basin at about 20f.with a few soundings up to 24f. Egmont atol l l ies betweenG reat Chagos reef and Pitt bank ; it is perfect in shape, 51 miles long by 2 broad ,

wit h onlyone passage to the nort h

,six islands to t he south , lagoon 10f. deep. S. by E. from G reat

Chagos lies Diego G arcia, anot her at oll , completely surrounded by land to the sout h,but

with three passages (35 , 31 , and 6 f. of water) to the north ; i t is 13 miles long by 6 milesin great est breadt h

,lagoon maximum depth 19 f , much Shoaler to the sout h. This atol l

was visi t ed in 18 8 5 by Mr G. C. B ourne, who subsequently gave an accoun t of i t . He camet o t he conclusion t hat the whole at ol l had been raised a few feet ‘.

The G reat Chagos bank is of somewhat irregular shape, about 8 2 miles east and westby 70 miles along a line a t righ t angles. I t has one island to t he north

,described as rocky

,

and seven to the west. The reef can nowhere be described as awash,save immediat ely in

the neighbourhood of some of t he islands, but the line al l round the bank is markedlyShal low , forming a reef. The general dep t h of the lat ter west and nort h is about 6 f.,least 4 f ,

while east and south it is about 10f., l east 6 f. The lagoon basin has an averagedepth of about 44 fi, greatest depth in cent re 48 f.; i t is connected t o t he sea by t wen tyt hree passages. The shal low

,encircling reefs are bet ween t hepassages

,great broad flat s

,sloping

in,

from t heir outer edges to a depth of 20f.; some are Obviously smal l at ol ls, havingshallower wat er a t the edges and deeper in t he cen t re. Against t he lagoon basin t he20

,

30 and 40 f l ines are practical ly con t erminous , t he reefs down t o 40 f. being as prec i pi t ous in t heir slope on this side as towards t he deep sea. About sixt een shoals arisein the lagoon from the bot t om in 40f to within 6 to 10f. of the surface , a st riking uniformity in dept h wit h t he rim. Alt hough the atol l from t he soundings would appear tohave been wel l surveyed i t is remarkable that there are in the lagoon no shoals between10 and 40f. in depth. Of the passages five have abou t 40f. of wat er right through , whilesix others appear t o have over 30f.

Nort h of G reat Chagos the Vict ory bank , 4 miles long by 21} broad , has a perfect r im

at 3 f. wit h 18 f. of water in the cen t re. Peros Banhos is a perfect atol l wi t h 27 is lands,

in shape nearly a square 13 miles across ; t he greater part of i t s reef is awash save t o

the S E , where it is submerged by abou t 4 f. of wat er , but even here two isle t s rise abovethe surface. The lagoon has thirteen passages to the exterior and is much studded wi t hShoals ; i t s general depth is less t han 30f ,

but one sounding of 41 f. is recorded. The wholeatol l bears a very close comparison to some

1 P.R.S ., vol.43, pp.440—461, 1888. M r Bournehasinformedmet ha t masses of theconglomerate(or reef- rock)ex is t everywhereon theou ter reef in pinna cles and bu t tresses ,ex tending out from the

“ bea ch between t ide-ma rks. The

s t rat ificat ion of thesei s s t a ted to behori zon ta l. No coralrock i s built up t hus abovethelow t idel imit , so tha t thesea bsolutely provea changeof level. M r Bournein h i s a ccoun tbases h is views ra ther on thehor i zon ta l s tra t ificat ion of

sh inglerock , ” an agglomera t ion of broken cora ls , molluscandech inoderm shel ls , etc. th is rock w as probably or igina l lya format ion on thelagoon s ideof a reef awash , and i t wouldbeinteres t ing to knowmoreof i t s dis t r ibu t ion. Theobservat ions on Ocyp odeI cannot admit as referring t o a cra b of th isgenus at all ; as M r Borrada i lepoint s out in hi s paper in

o f the more sout hern ones in t he Maldives.

th is publ ica t ion the crabs of th is genus do not const ructburrow s l ined with weed

,and oneSpecies l ives abovet ida l

l imit s ; the holes of shore forms cave in, and aremade

a fresh a fter every r ise of t ide, not necessari ly too in the

sameplace. Mr Bourne’s deduct ions on lagoon forma t ioni n genera l by solut ion do not seem to me to carry muchweigh t. They are based on ex per ienceof onea tol l a loneand tha t of abnormal form, i n that i t i s of considerables i zebut completely surrounded w i t h land ex cept to thenorth ;

t hecirculat ion of wa ter in the sou th ha lf of the lagoon ,

wh ich i s especial ly referred t o, cannot be any th ing but

incons iderableas compared even to any a tol l with reef onlyawa sh in severa l pos i t ions.

CORAL REEFS OF THE INDIAN OCEAN. 19

Salomon at ol l t o t he east of the last is oval , 5 miles long by 3 broad , wi t h eleven islandson the rim

,which is awash everywhere save for a single passage to t he nort h ,

which i t selfhas a shoal of 1 f. in the cent re ; t he lagoon averages 12 f deep , great est sounding 17 f.B lenheim reef to t he E.N.E. of the last, 6 miles long by 2 broad

,would closely resemble

Diego G arcia,i f the lat ter instead of land had on ly a reef

,swept by t he tides ; i t s open

ing,however

,lies to the south - west , and i t s lagoon has only 10f. of wat er. Final ly ,

Speaker’s Bank to the nort h of th e last two is an irregular oval reef,24 miles long by

13 broad ; dept hs of 5 t o 7 f are recorded righ t round its circumferen t ial reef wit h 24 f. inthe central basin.

4. MALDIVE SECTION.As my remarks on the Maldives are mainly based on my own observations and soundings

I Shall defer t hem ,t oget her wi t h t hose I may have to offer on t he Laccadives , to a

separat e chapt er. The most sou t hern at oll of the Maldive group,Addu

,lies due north of

the Chagos Archipelago at a dis t ance of 240 miles,and the lat ter group migh t wel l be

considered to be a cont inuation of t he Laccadive—Maldive line. There is , however, onesounding of 2500f. so close t o the centre of the channel between

,t ha t i t i s pract ical ly

certain that t here is no bank at abou t 1000 f , connecting the Chagos , as assumed byMr Bourne. Since to the nort h the Laccadive banks very closely approach t he Indiancoast

,it is important t o note t hat t he 1000f. line has been traced everywhere between.The Indian Peninsula has pract ical ly no reefs of any sort ; its southern part t owards

Ceylon partakes of t he same formation as the north port ion of that island. The latter haslocal ly fringi ng reefs , mostly in bays along t he sou t h and north coasts ; above these to t he

sou t h a loose rock, composed mainly Of fragments of coral, often forms broad flats,3 to 4 feet

above high tide,now being rapid ly washed away wherever they impinge on t he beach 1

.

Near Dambula in t he cent re of t he island I saw limestone hi l ls , at least 700 feet high ,which from t heir general cont our I Should j udge to have been of reef- formation ; the rockis now completely crystallised , al l trace of organic structure being lost. The JaffnaPeninsula , the outlying is lands and the nort hern part of Ceylon are al l formed of raisedcoral line rock

,which st i l l retains most of i t s organic st ruct ure ; t he elevation , as I saw it,

did not appear to me to have been more t han 40 feet , i f as much. A former land connec t ion w it h India was clearly indicated , of which a reef, known as Adam ’s Bridge

,between

the islands of Manar and Ramesvaram,is the rema ins z.

1 Thesecan beseen in numerous places between Mata le Government Museum, B ullet in 3 , 1895 ; a lso Repor t B ri t.

and Wel igama. Ass., p.400, 1900.2 See Ramesvaram Is land,” by Edgar Thurs ton, Madras

CHAPTER II.

METEOROLOGICAL CONDITIONS.

THE Challenger Reports , edit ed by Sir John Murray, have shown how important it ist o consider t he meteorological conditions of any region whose geography it is desired tostudy. In the “ Int roduct ion t o t his work I have briefly indicated t he great variabi li t yfound in the topography of the different at olls of the Maldives. As the latt er grouptoge t her with t he Laccadives extends from lat. 1

° S. to la t . 14° N it wil l be readily

understood t hat t hemet eorological conditions must vary very widely over t his great region.

In respect to coral reefs,winds and st orms are usual ly supposed t o be the chief factors

in t he formation of land. Dana 1 has point ed out that their distribu t ion fol lows the isocrymeof 68 ° F. both N. and S. of the equator. Semper 2 drew att ention to the effects of curren t sin shaping reefs, a point which has subsequently been furt her emphasized by several naturalis t s. The eroding act ion of rain has been referred t o by many au t hors

,chief among

whom is Prof. Agassiz , who has also laid great s t ress on t he action of the sea and waveson al l coast s in general

,but on coral - Shores in particu lar". The subj ect t hen being of wide

interest in respect to t he format ion of coral reefs and islands, I shal l consider t hemeteorologica l condi t ions of t he regi on wi t hou t in this place examining minu t ely their effect s.The following account is derived largely from personal observat ions and native information

,

wi t h which I endeavoured to check the account in the “ West Coast of H indust an Pi lotI have also consulted t he captains of various st eamers , which regularly ply in these

waters.

The W inds, e x perienced in the Laccadives and Maldives, are t hose of t he two monsoons,

influenced to t he north by t he prox imity of t he Indian coast and affected in the sou t hby the equat orial winds. The nort h - east monsoon becomes set in the Laccadives abou tt he end of November and cont inues unt i l t he end of March ; during this period a more orless northerly wind prevai ls toget her wit h long calms

,but litt le or no heavy weat her. The

monsoon gradually t ravels down t he Maldive group, becoming regu lar from N.N.E t o E.N.E.

in Male at ol l early in January. I t is ushered in by a fortnight of st rong winds fromnort h to east with heavy rain squalls. The monsoon becomes definit ely set as far sou t has Suvadiva atol l before t he end of January. It is felt too even in Addu (la t . 0

°

40' S.)

1 Cora ls and Cora l I slands , 3rded.pp.108 and 335. 3 Vi de TheIs lands and Cora l Reefs ofF ij i (Bull.Mus.

2 An imal Life, ” Interna t. S ci . Ser ies , chaps. VI I. and Comp . Z 001. vol. x x x I I I. pp. 1—167 , 1899) and numerousVIII. previous publ icat ions.

M ETEOROLOGICAL CONDITIONS. 21

from January t o March , but the wind is here very irregu lar, veering at any t ime sixpoints t o eit her side of nort h - east. Heavy rain squalls are frequent during t he whole ofthe monsoon in t he sou t hern atol ls, but the wind , alt hough oft en blowing very fresh , haslit tle weight. Fine weather, little rain and calms may be expected at t he same time tothe nort h of Male at oll.

The south - west monsoon is not felt in Addu atoll,where the “ Pi lot s t ates that from

May to December t he winds are from west to south and S.S.E. wit h much rain and squalls.It is wel l set in the Maldives during the mon t hs of June , July and August ; fresh galeswit h heavy rain squalls occur frequen t ly from any direction within four points of sou t hwest. The period in t he Laccadives is rather longer

, t he south - west wind usually becomingdefinitely set t owards t he end of May and continuing regu larly unti l September. As alreadymen t ioned in the “ Introduction i t did not in 1899 commence at M inikoi unti l t he secondweek in August

,up to which time t heprevai ling winds were from west to W.N.W.

The months,which int ervene between the monsoons

,form in the Maldives a period

during which winds from any direct ion may be fel t. Fresh gales of short duration t endt o come up , the wind commonly veering from the north or south

,right round to eas t , or

west. In Apri l we experienced in t he channels between Addu and Suvadiva ,and between

the latter and Hadduma t i , heavy gales from t he west. During t hese intervening periodscyclon ic st orms or hurricanes are especially liable to occur in the Arabian Sea. I haveobtained no record of w inds of this force having been ever experienced further south t hanMale at ol l. They occasionally sweep over Ti ladumma t i and Ihavandifulu atol ls

,the las t

occasion remembered by t he natives being in February abou t 24 years ago, when verygreat destruct ion is st ated t o have been wrought among t he larger native vessels , manyof w hich were at t he t ime lying at anchor in the lagoons. As negat ive evidence of theabsence of such st rong winds I may refer t o the fact t hat throughout t he Maldives Ihave found no true negro - heads—actual part s of the reef—which have been thrown upby the breakers on t o t he reeffla t. Hurricanes visit Minikoi at in t ervals of about 12 years,but

,their cent res usual ly lying considerably t o t he north

,t hey may pract ically be neglected

as fact ors in Shaping t he original reef and land. The more nort hern part s of t he Lac

cadives are,however, much affected. The effect s of hurricanes in this locality are admirably

described in the following account of one , which visited Kalpeni in 1847 The sea washedover t he island, and the st orm - wave carried away the very soi l of t he narrow er part s. Of

1600 inhabitant s 250 were washed away and drowned in that nigh t , 100 t o 15 0 died offamine and disease

,abou t 450 remain on the island and the rest have dispersed t hem

selves. Of a lakh of (coconut) t rees abou t 700 sickly ones are st anding 1.”

At

Androt of a popu lat ion of over 3000 only 900 were left aft er the same hurricane.

The c u rren t s are of course largely dependen t on the winds and tide. The direct ionfrom November to March in t he Maldives t o t he north of la t . 2° N. varied from nort h ~

west t o south - west ; t he drift in the cen t re of t hegroup averaged about 18 miles per diemduring t he same t ime. Alt hough we experienced st rong winds from the west during Aprilin the “ Equat orial ” and the “ One- and - a - half- Degree channels

,there was a t the same

time pract ical ly no current. The nat ive Skippers , however, will not s tart across eitherchannel un less the wind is fairly st rong and a t least nine points off the direct line from

1 “ Report on theLaccadiveIslands , da ted 19 May , 1848 , by W .Robinson, Esq., Madras, p.76, 1874.


at ol l t o at ol l ; they t hen set a course abou t 12 miles t o east or west of the at oll t heysubsequent ly hope t o reach. Wi t h t he sou t h - west monsoon comes a change of current

,

which is ex perienced a t Male about t he midd le of May,t he set t hen being almost due

east ; t he rat e of t he current varies wi t h the wind of the previous day. During Apriland Oct ober little or no current is fe lt .

The curren t s in t he northern part of the Laccadives are very irregular owing to theproximity of t he Indian coas t . This is especially t he case in the sou t h - west monsoon

,a

definite sou t h—eastern set being often ex perienced almost reaching to Minikoi. However,

the following dat a are given in t he Admiralty Charts for the neighbourhood of t he lat t erat oll

J anuary 0—4 5 mi les per 24 hours , N.W.

February Various,any direc t ion.

March Various,any direct ion, often dueN. or S.

Apri l 10—25 mi les per 24 hours, S.varying 4 point s t o E. or W .

May Various,a lways sout herly.

0—45 mi les per 24 hours , S E.J uly 5—30 S

August 0—45 E.S.E.

September 0—25 S.E.

Oct ober 10—25 E

November Various, 0—35 mi les per 24 hour s, W .N.W .

December 10—25 N.N.W .

The maximum currents above are undoubt edly current ! wind. Many captains of t he

gr eat passenger liners, running t o t he east, do not make any allowance, based on the chart s,but consider only t he winds of the previou s days in set t ing their courses. The “ WestCoast of H indust an Pilot ” makes t he followi ng s t at ement (p. 363) in respect t o the

“ Eigh tDegrees Channel between M inikoi and t he Maldives z—“ I t will be prudent to keep nearert o M inikoi t han t o t he Maldives

,as t he current sets to southw ard - between the end of

Sept ember and t he close of the year. In t he strengt h of t he sou t h—w est monsoon , midchannel is t he best t rack , or rat her inclini ng to t he Maldives. I t is scarcely necessaryfor me to point out , t hat t his stat emen t is seriously , in confl ict wi th t he data of t he

chart s. It serves excel lent ly t o Show t hat t here are no definite curren t s in the Laccadives,

al l S ides of t he at ol ls and reefs being probably washed equal ly a t differen t seasons of theyear.

The curren t s t hrough t he atolls in the Maldives are of great strengt h,as also in the

channels between the at ol ls. A current of t hree to four knots per hour is by no meansuncommon in t he passages on t he east side of t he atol ls in the north - east monsoon, andt his was even exceeded in t he channels t hrough the leeward reefs. In t he lagoon ofM iladumadulu we experienced in the midd le of December a current of 212

~ to 3 miles

per hour, against which we could make no headway. Wit h st rong easterly breezes—henceat least eigh t points off our t rack in both at ol ls—we made a course along the east sidesof N. Mahlos and Miladumadulu ,

which,if accurately plott ed

,would Show a series of

loops,losing opposite the passages and gaining our cas t ing again t o leeward of t he islands

and reefs. To test t he rat e of t he curren t s, I adapted Messrs Negretti and Zambra’s

river—meter by fi t t ing i t wit h a larger vane. Wit h currents over 2 kn ot s per hour itappeared t o give fairly reliable resul t s. I took a series of observations on t he currents

M ETEOROLOGICAL CONDITI ONS. 23

near Male island during the third week of February , 1900, wind 2 to 4 Admiral ty scale,

regular never varying more t han a point from north - eas t , dep t h 2 fa t homs.

Direct ion Current in mi les per hourPos i t ion of

Curren t Greatest Leas t

PassageMa let o H ululeW ilingi li

Funadu

Tw o miles dueN.ofMa leChannel N. t o S.Ma leAtoll

0 0 0 0 0 0 0 0 0 0 0 0

I also sai led t wice up t o the passage be tween Hulule faro and Furenafuri , finding current sof 3 7 and 2 1 knots in t o the lagoon. Of course al l these current s were largely dependentupon t he tide , but it is a most noticeable feature t hat t he direct ion remained constant inspite of the latter, which hence on ly affected the rat e.

The depth , to which the curren t s of t he Indian Ocean extend,has never

,so far as

I am aw are,been investigated. Yet in considering any t heory of the format ion of coral

reefs i t is a most important factor,requiring consideration. I may perhaps refer t o a

Short paper by Admiral Sir W.Wharton on t he “ Foundat ions of Coral Atol ls 2,” where he

discusses in this connect ion t he dept h t o which there may be motion sufficiently violen tto move material.

I applied t w o methods of investigat ion to ascert ain the directions of the currents ,i .e. by examining t he specific gravity and temperature of the water at various depths, andby means of a direct ion - indicat or. Admiral Marakoff is t he chief exponent of t he firstme thod

,w hich he used to examine

—

t he curren t s of t he B osphorus 3 ; t he direction of t he

current is found purely by deduction“. I employed it in the channels round Male, ment ioned above , but t he differences were t oo smal l and variable for it alone to be a sat isfact orycriterion. I t was usefu l merely as a check on my second met hod. In t he latt er I attempted to record t he direct ions of t he currents at any depth by means of a compass

(made for me by Messrs Negre t t i and Zambra). The lat t er was quit e open to t he wat erat t he side of i t s box ,

and could at any dept h be jammed by a messenger. The wholewas placed in the middle of a V- shaped vane

,which swung freely on a cen t ral bea ring,

it sel f clamped on the sounding line. Immediately below t he direction - indicator I placedmy curren t - meter, which I fitted to swing freely

,and at the end of t he rope I at tached

a 20 lb. sounding lead. The compass gave uni form resu lts , but the readings of t he curren tme t er varied so greatly t hat I omit any further reference to them.

I carr ied out observat ions in t he fi rst place t o ascertain,whether t he curren t s were

the same a t al l dep t hs in t he deeper passages of at olls , as from t heoretical considera t ionsI expect ed to find under- current s. I found as a resu lt t hat there was no difference in

1 Approx ima teon ly. I obt a ined my sca leon themeter 3 I had the pleasure of discuss ing th is subjec t w i thby means of t imed floa t s , u sing for the purpose corkedbot t les , weigh ted wi th sand, w i th t iny flags , so tha t theyshould not be affected by thewind. I t mus t be checkedfrequen t ly as the amount of frict ion may a l ter grea t ly.Previous ly to th is I h ad used themeter const an t ly in M inikoilagoon.

2 Na ture, vol.Lv.pp.390—3 , 1897.

Adm. S ir W .W harton,S ir John M urray and M r J . Y .

B uchanan ,beforeI left England. I am indebted to these

gen t lemen for much pract ica l advi ceas to thework of t heex pedi t ion. I am only sorry that I could not carry out theirsugges t ions in their en t irety.

4 Vi de LeVi t ia z et l ’ocean Pac ifique.


any Of the posi t ions previously recorded round Male island , eit her at half- flood, or halfebb from the surface t o within 3 fat homs Of the bottom

,an Observat ion I confirmed by

the hydrometer. In the chann el between N. and S. Male at ol ls t he current was in thesame direction down t o 1 50 fat homs.

My second series of observat ions were to ascertain t he dept h to which t he mainoceanic current s extended and their behaviour on meeting w it h an Obstacle. I made twoexaminations

, t he first at M inikoi and t he second at Male‘. A preliminary series ofObservat ions Off the east side of M inikoi showed that the oceanic current was sett ingS.E. by E., and cou ld be recorded usually down to a dept h Of at least 60 fat homs. Start ingfrom Weli Gandola passage, I then ran a line Of soundings ou t from the reef, on whichI had affixed a pole. At t he same t ime I t ook a series of Observations on the directionOf t he current at the surface and at intervals Of about 20 fat homs down to a depth of120 fat homs. Fig. 5 shows my result s graphical ly ; it records my stat ions and t he vertical

S EA LEVEL

F IG. 5 . Diagram to i llust ra tethebeh avi our of a current on s t r iking a reef. The sect ion i s that of the outer slopeOfi Wel i - Gandola channel , M inikoi. Hori zon ta l and vert ical scales the same; the former is marked in yards,

t he la t ter in fa thoms.

effects of t he Obst acle on t he current. Horizontal plans would show the divaricat ionOf t he upper current round t he atol l , and t he gradual merging int o this Of the lowercurrent. I made two addi t ional soundings betw een t he reef and the first recorded inFig. 5 , but the rollers were so heavy t hat I did not at t empt to use the indicator. Asimilar line

,run Off the reef Of Hulule faro , t he nort h point Of Hu lule bearing due west ,

gave similar result s down to 80 fat homs , 500 yards from t he edge Of the reef. The recordsthough were complicated by t he prox imi ty of t he deep channel between N. and S. Maleatolls.

1 An apology i s perhaps needed for offer ing theseobserva apparat us. I proposed subsequent ly to invest igatethepointt ions a t a ll ; they are not in tended to befinal , but may fully in theMa ldives

,but I w as unableon account of i l lness

perhaps suggest t o others,morehappily si t ua ted than I am, to makemorethan a prel iminary series of observa t ions 03

theneed of invest iga t ing thema t ter fully. My work , indeed, H ulule.at M in ikoi w as meant t o be ex per imen ta l to tes t my



nigh t or day ; such temperat ure is, however, very except ional. The wet and dry bulbs arenever far apart

,and t hey commonly approach wit hin 1

° of one another during t he sout hwest monsoon. The t empera ture Of t he sea varies great ly from year t o year. The Admiral tyChart s for t he region Of Minikoi give February May August 81° and November

My Observat ions for t he four mon t hs I spent in t his at ol l were,expressed in round

numbers,June July August 83

° and Sep tember The maximum and minimumdaily t emperatures seldom varied more t han 4

°

a t dept hs Of 4 to 8 feet ; t he Observat ions,recorded above

,are the mean of t hese. The temperat ure ou t side t he reef and w it hin the

lagoon never varied more t han 2° F.

[Nor a Mr Forster Cooper on reading over t he above chapter has Offered me t he followingObserva t ions in respect to his cru ise in March and Apri l

,1900 In my jou rney t hrough t he

a t olls Of Sou t h Ma le, Felidu ,Mu laku

,Kolumadulu and H adduma t i I ex per ienced only north

westerly winds wi th t he ex cept ion of a due easterly ga le, last ing t hreedays, while I w as in t he

sout h of Felidu a t oll (Ma rch 19

“ Li t t le current w as not iceable during t he c ru ise wi t h in t heencircl ing reefs Of t he a t olls, butin t he div iding channels i t var ied us ua lly bet ween 2§ and 3 knot s per hour, set t ing westerly. Int he channel between Kolumadulu and H adduma t i of Veimandu island t he c u rrent reached 65 knot sper

‘

hour. I t appears in t he nort h -east monsoon t o be a lways very s t rong in t h is channel ; a na t iveboa t w as a few years ago carried completely away, event ua l ly reachi ng t he African coast nearZ anz ibar.

The current s Of ei t her monsoon tend to run very s t rongly up and down t he Ma ldive cha in.As above shown, t hey reach t o a cons iderable dept h, and na t ura lly tend t o be much impeded by

t he banks. A st rong cu rrent wou ld be ex pected in the Veimandu channel,t he NE . horn of

H addumat i direct ing t hedown c urrent in t h is direc t ion.Capt . Pigot t , S.S. “ Vasna

, CO., informs me t ha t on t he n igh t Of J une25 t h,1899

,he ex perienced a set S.5 1

°

E. of 345 mi les in 12§ hou rs Off t he eas t Of M iladumadulu

and Ti laduma t i a t olls. The wea t her w as fine and set t led,wi th a l igh t wes terly breeze, t he sea

ca lm wit h a long low swell from t he sout h -east. The cou rse w as M afarO a t 420p.m.

bear ing 12 mi les due west . In t he middle of the “ E igh t and a H alf Degrees Channel ’t he

cu rrent w a s set t ing abou t 2 knot per hour due E. Th is Observa t ion shows very admi rably how

t hecurrent washes around t hebanks.)

CHAPTER III.

THE ATOLL OF M INIKOI.

SECTION 1. GENERAL.

M ini koi at oll (Fig. 1) is Of a more or less oval shape,somewhat pointed t o the north

east, where there is a narrow channel in t o t he lagoon wit h 2 to 25 fathoms Of wat er.This passage is fur ther broken by a shoal in t he cen t re

,showing clearly t he continuation

of the reef. The lat ter is for the rest perfect, and can i n moderate weat her be traversedeverywhere on foot at low spring - t ides. Three boat - passages exist to t he nort h - nort h - west ,avai lable a t half t ide

, but t he import ance Of these must not be magnified,as t hey are

regularly cleared by t he inhabitants. The great est length of the at oll is 5 miles andbreadt h 2 9 miles ; the circumference is nearly 13 miles and the lagoon occupies an areaof abou t 6 square miles.

The position of the at ol l has already been seen. It suffices here merely to drawat t ention t o i t s isolat ion , arising as it does probably within two miles from a“ depth Of

1 100 fat homs on every side except to the sout h - south - west. In the latter position a ridgeat a dept h Of about 950 fat homs appears from the charts to connect M inikoi at ol l t o t he

Maldive group. Unfortunately no proper sections were run when the atol l was re- surveyedin 1891 , and I had not the means to make a proper series of soundings. Indeed t he

weat her was t oo had during t he great er part of my st ay for boat s t o get out except t o

the north and east. In these situat ions t he slope commences wi t h a sudden drop from t he

reef t o 2 or 3 fa thoms , the bottom consisting Of rounded bu t tresses and - masses Of coralrock. Local variat ions due t o shal low plat forms occur

,but t he fal l is general ly fairly even

for 100 to 350 yards, where a steep commences at a dept h Of 30 t o 50 fathoms. TO

the east - nort h - east Of Kodi point t his shel f is more e x t ensive, t he greater part havinga dep t h Of 35 fat homs

,forming a possible anchorage in t he sout h - west monsoon. The

furt her slope here is more gradual,and the steep less defined , indicating t he t rend of t he

submarine elevat ion,on which t he atoll is founded. The steep is probably similar to that

Of Funafuti and other atolls,a few yards giving no bott om at 120 fat homs (Fig.

I did not sound deeper, but the chart indicates that a more gradual slope soon commencesdown to the encircling depth. The first slope to abou t 40 fat homs is notably broaderto the nort h and west than to the south and east ; in the lat t er posit ion no two sect ionsare t he same. In Fig. 5 a section Off Weli Gandola passage is represented. The bott om

4—2


w ould appear to be extremely rough , great masses of rock , off which the lead drops sometimes for an addit ional 3 to 5 fathoms.

The chief island,cal led M inikoi throughout t his paper, extends from the north point

right down the east and south sides of the atoll for a distance of six miles. It is onlyabout 50 yards broad to the north , but gradually increases further sou t h , attaining at itscentre a breadth Of 7 50 yards. In t he latter position a prominen t point , Mon - Rambu ,

may Pomf

era tu Point

FIG. 6. M inikoi Atoll (sligh t ly a ltered from the Admi ra lty Char t ). The rocky area of the land i s shaded, the

boulder zone dot ted, and the 100 fa thom l ine i s shown. A—G ,pos it ions of thesect ions in Fig. 7.

ex t ends to t he sou t h - east , and on the opposit e lagoon—shore t he only vil lage Of the atol lis situat ed. Proceeding sou t hwards M inikoi varies in bread t h between 300 and 600 yards ,ending wit h a smal l horn ,

ext ending nort h for 200 yards into the lagoon. This horn ist he only bank in t he at ol l , which may be supposed to have been formed by sand anddet rit us swept over t he reef.

On t he seaward face from t he nort h end t o Mon - Rambu point the reef is very narrow,

and indeed can scarcely be said to exist. It t hen gradually broadens to the west pointof M inikoi island , where i t passes int o t he reeffla t Of t he west and nort h sides Of t he

a tol l. The island i t sel f merges in t o a wel l - marked boulder zone on t he same reef, onw hich t he small island of W iringili arises at a distance Of half a mi le. The reef ret ains

THE ATOLL OF M INIKOI. 29

a characterist ic broad reeffla t and rough zone for one and a half mi les to a smal l rockypatch

,B agandi island , extending from t hence in an east - north - east erly direc t ion to t he deep

lagoon passage. In t he latt er part the reef is much na rrower and considerably broken up,the boulder zone scarcely exist ing in some places.

SECTION 2. LAND.The surface of M inikoi island (Fig. 6) is sharply divided int o t wo areas , an outer,

covered on t he surface with large , loose, coral or rock masses,and an inner

,with sand.

The line of j unct ion is qui t e distinct , in some places a slight dip,in others an abrupt

step down to t he sandy area. The rocky area fringes the island on its seaward face ,varying up to 1 50 yards in bread t h near the vi llage. Where the island is less t han 100

yards in breadt h,it alone is traversed in a section. The whole of the nort hern two miles

thus belongs to this area wit h t he exception of a narrow sandy pat ch against the lagoonat Boni - Kodi. Sou t h of Mon - Rambu point t he rocky area narrows somewhat

,especial ly in

KO- Vari Bay and between Rocbera and Tevera t u points,where it is in places only 70 to

80 yards broad , being indeed general ly less than 100 yards.

In the rocky area t he surface is covered with masses of coral or limestone,some

round,others flat slabs , weighing up to l cw t . or even more. To the north of Mon - Rambu

they are much larger than to the south,where blown sand also has t o some e x ten t inter

mingled. All the surface masses are pi t ted , and eat en into sharp points by the rain , whichrapidly drains t hrough t hem. NO soi l has formed , and layer aft er layer Of blocks can beremoved

,unti l a firm bed conglomerat e within tidal limi t s is at last reached. The whole

is covered wit h t imber, or woody herbage , the roots of which extend down even into theconglomerate. On the surface the masses have much t he same appearance , allowing forrain

,as t hey have on a tidal beach. If the outer blocks however be removed , M adrep ora ,

Poci llopor a ,and ot her branching corals are found wit h t heir stems sti l l unbroken , while

massive species have their calicles and septa even yet entire,

absolu tely nega ti ving the

p ossibi li ty of a beach- or igi n for therocky a rea.

Speaking quite genera l ly,t he island at tains i ts maximum height t oward the seaward

side,and slopes gradual ly to the lagoon. To ascertain t he height accurat ely, I ran six

sect ions across t he island , leaving one broad gap bet ween the ligh t house and t he vil lage ,w here t he growth Of coconu t and other timber was too dense to allow Of rides being cut.

Even there , however , I Obtained a few partial sect ions in from the seaward beach. Theinner part Of the reef- fia t , where t he latt er exists , lies a t the mean low - t ide level, andal l heigh t s are t aken as above this

,since i t must be regarded as t he upward limit, t o

which the various organisms can“ build their reef. The sec t ions give the fol lowing measurements (Figs. 6 and

Breadth of Is land Heigh t a t Cent reSect i on m yards of Sect ion370

495

37 5

103

80

107


In most Of the sections the greatest height is close to t he t op of the seaward beach ,but in F i t l ies 66 feet back from this posi t ion. A sevent h section G (Fig. Oppositethe vi l lage in Ko- Vari bay, gave an elevation Of 19 ft. 3 in., close to the sea

,and heigh t s

up t o 20 fee t occur on bot h sides of Mon - Rambu point,decreasing northwards to a general

heigh t Of 14 feet and to the sou t h Of 125 feet. The abnormal heigh t Of the beach continuesto Tevera t u point right along KO- Vari bay. At my first examinat ion I did not deem thisexcept ional , since on t his part

,especially supposing the atol l at one time a reef awash ,

F IG. 7. Sect ions across M inikoi (seeFig. In every case the seaw ard face i s represented to ther igh t. Ver t i ca lsca le8 t imes t hehor i z onta l.

masses Of rock and coral wou ld , I considered , natural ly tend to be piled up in each monsoon. Inside t his ridge t he surface of the land had been ,

I saw , much lowered by previousgenerat ions of t he nat ives for planting purposes

,and also to form pits for macerating the

coconut husks for coir ‘

My attention being especial ly draw n to t his,I re- examined the whole careful ly, and

dug a number of pit s , coming t o t he conclusion t hat the ridge may have been broadenedby t he above means

,but that its height is due to absolutely nat ural causes. An eleva tion

then, a llow ing for ra inw a ter denuda ti on, of a t leas t 24 feet demands exp la na t ion.

1 Th is hol low ing out ofgarden land—totam, as i t is called I did not meet with i t inany part of theMaldives. Coir i s—is common throughout thewholeof theLaccadives for ropemadefrom thefibreof thecoconut husk.ragghee (gra in) growing, and also for banana s and yams.

F ig . I . Sea - beac h a t low t idenea r Kodi poin t , M in ikoi , w it h Th runs coe a shore

F ig. 2. L agoon - bea ch a t low t idenea r Kodi point , M in ikoiToface15.31



broadening sligh t ly. Southwards again t he rock reassert s itsel f wi t h smal ler coral massesand a lower cliff (seldom more than 6 feet ) wit h perhaps a t endency to form a greater t alus.The basal conglomerate masses too are generally smal ler and Of lesser heigh t . Coral slabsfrom t he lagoon do not add a t al l t o t he beach

,and even seaweeds and sand do not

accum ulate excep t in a small bay near Kodi point,protect ed by spurs Of rock on ei ther

side. The shore especial ly a t

Boni - Kodi is st rew n withfallen coconut and other t imber. An old pilgrimage pat h

,

extending t o t he shrine of aS a id, closeto Kodi point, onlye x ist s now here and t here.The end wal l of a long house

,

which formerly st ood a t Bal iMi - B ungi, t he leper settlemen t , alone st ands. Thesea

i s r ap idly rega i ning i ts sw ay.

Along the sand area thecliff ret ains i t s posi t ion almostdown t o t hevi l lage

,and the

beach is st rewn wit h coconuttrees. G radual ly however t heslope broadens , and i t s summitOpposi t e thevil lage is almostF IG. 8. Clifi aga i nst the lagoon near Kodi pomt , showmg conglomerate h d I) ll d

.

masses a t base. (From a p hotograp h).was 6 y t e 111gwaves Ofthe h i gh spri ng t i des.

At Moli - Mati point i t s bread t h reaches a maximum ,50 yards across bet ween t ide- marks

,

the land above gradual ly attaining a heigh t Of 2 feet more. The high - t ide limit is shownin t he sou t h - west monsoon by a line Of green algae and sand washed up by the waves.Shel ls in the beach are rare

,t he whole being formed of extremely fine, almost impalpable

whi t e sand , t he washings of many reefs. NO decaying vegetable mat ter is seen t o relievet he whit eness , save t he single line left by t he preceding high t ide , al l such being quicklyseized by Ocypod crabs and drawn in t o t heir burrows.

Above the high - tide limit the height of t he land is gradual ly increased by sand blow nup t he beach a t low tide. A few yards back herbaceous plan t s commence to grow

,soon

giving place to the t ypical bushes. The nat ives quickly fol low , and plant coconuts , whichmay be seen gradually rising in heigh t furt her inland. The land, indeed, i s here r apidly

ga ining on the lagoon. S t i l l furt her on , t owards the inner side Of Tunda point,t he cliff

and fal len trees reappea r, where , i f anywhere , i t would be most nat ural t o expect abackwash and a piling up of sand t o form fresh land.

Here in M inikoi we have a typically sit uated island , of an atol l wit h t he reef continuingunbroken a t i t s west end

,round which t he t ide surges wi t h considerable force. The

current should sweep sand and mud Off t he reef, should bui ld up a shoal int o t he

lagoon , and shou ld carry the poin t of t he island along t his. It is essen t ial t o see if t his

THE ATOLL OF M INIKOI . 33

is what is real ly happening, since Dr G uppy l has brought forward a great mass of circums t ant ial evidence to show that i t is in t his way t hat secondary a tollons are formedreclaiming ult imately large areas from t he sea. The sou t hern half Of M inikoi too wouldseem to be absolu t ely similar to t hat Of t he Cocos - Keeling atol l described by G uppy. Therocky seaward rim Of M inikoi ex t ends a t Rochera point direct ly into t he boulder zoneOf t he reef, many long masses Of rock more or less undermined marking its transit ion.The rocky area before this had from t he light house westward been gradual ly narrowing

,

and a t the point t he beach has a smal l cliff above with undermined or fal len bushes,

which conclusively prove i t s washing away. Jus t behind Rocbera is a smal l marsh orkuli

”

,covered wi t h reed , its floor below the high - tide limit. Tunda point does not e x t end

direct ly into the lagoon , but a littl e to t he nor th - wes t. Round it t he wat er is deeperclose to t he shore t han furt her out t owards W iringi li , and t he great er part of t he bot t omhere is covered w ith a branching species Of Por i tes

, right t o t he base Of the beach,some

being thrown up at the point. Nort h of t he kuli al l is sand , the same indeed as formst he lagoon beach a t Moli - Mat i poin t . If some Of t he beach sand be tween Rocbera andTunda points (not a mere surface scraping) be sifted to remove the smallest part icles , itwil l be seen - t o be essential ly the same. The sand carried over the reef in this posit ionis largely foramini feral , while t hat on t he beach is formed Of finely tri t urat ed coral part icles. I t follow s then tha t Tunda p oint must or igi na lly have been a lagoon forma t i on.Behind this point on each side the shores show a marked loss. The la nd i s gradually

nar row ing , the sea w i ll sooner or la ter break a cross , the cor a ls round thepoi nt w i ll besi lted

up and ki lled fina lly the w hole w ill be w ashed a way. M oli—Ma t i w i ll probably, unless man

interferes , form such another p oin t in thenot very di stant future.

B efore discussing t he formation Of the conglomerate , it is import ant t o consider theaction—as deduced from Observation—Of t he various agencies

,to which a reef- rock

,such as I

consider this rock t o be , may be subject ed above and below the sea. A reef, freely exposed tothe sea, may be compared in struct ure t o a sponge, a series Of hollow cavities j oined t ogetherby canals. Corals grow leaving cavit ies and spaces, which become si l t ed up t o some ex t entwi t h sand , composed Of their own fragment s and Foraminifera. These spaces are nex t bridgedover partial ly by means Of t he coral s themselves , but more part icu larly by Li thothamnion andPolytrema

3(Foraminifera). When the cavi t ies are once comple tely closed in from t he sea

by organic grow t hs , t here i s pract ically no circu lation of water in them ,causing any change;

boring and other organisms are ki l led,or depar t for t he more surface growt hs. As long

as t here is a circulat ion Of wat er,a cont inual solu t ion and redeposition Of t he l ime “ goes on.

The sand,being loose

,is ground down into fine particles , which nat urally, presen t ing a

relatively great er surface , would be first dissolved. The deposi t ion would tend to occurwit hin t he pores Of t he corals

,where the change Of water is less

,and t here is a fixed

support ; the lime deposi t ed fo l lows t he cryst al line form of the coral. SO long as the coralskeletons are protected by organic grow t hs , none Of t hese changes can go on

,but boring

1 TheCocos - Keel ing Islands , S cot.Geo.M ag. vol.43 , p.442,9 I propose to employ th is term for all such marshes 3 Th is form gradua l ly increases in impor tanceto a limi ted

of a toll isla nds , borrow ing thename from theMa ldivan, depth.whencei s a lso der ived thenamea tolu or a t ol l. Theterm 4 For the sakeOf brevi ty I use th is term throughou t.appears to me preferable to barachoi s , wh i ch i s u sed for instead of ca lcium carbona te, or bicarbonate.such by G.C.Bournein speaking ofDiego Garcia (P.R.S?G.

34 J. STANLEY GARDINER.organisms, especial ly sponges , may work destruc t ion.

_

If t he growth of the reef be vigorous ,t he whole action of the lat ter may be neglect ed

,t he corals not being seriously weakened.

If on t hecon t rary t he condit ions are unfavourable , whole masses of coral may col lapse witht he sligh t est superincumben t weight in t o smal l grains, giving, wit h t he solu t ion and redepos i

t ion Of t he l ime , great bricks , as i t were , of st ruct ureless rock. Ot her skelet ons suffer wi t ht hose of corals, but t he solvent act ion is invariably more and t he boring act ion less pronounced.

Above t he sea and t ide - marks the first action on the same reef- rock is the solutionby the rain of t he soft er parts, sand , nul lipore, et c.1 Some of t he water escapes carryingi t s l ime

, but a considerable portion percolat es int o hollows and is dried up, its lime servingto consolidat e the sand

,et c. The wat er att acks t he corals from the base

,where they are

weakened by boring organisms,breaking them off and causing a S inking of the whole

surface , port ion by port ion. This continues unti l t he crevices below are t o a large extentfi l led up wi t h sand and fragmen t s , more or less consolidat ed t oget her. There resu l t s final lya solid surface

,strewn wit h coral blocks, which have less rapidly dissolved. A few chambers

may remain in t he rock ; t hey ei t her exis t ed as such from the original reef, in which theywere closed in , or were subsequently dissolved out owing to the rain having ready meansOf access and escape.

B etween t ide - marks the act ion is more complicated by the wetting and part ial dryingOf the rocks by water

,contain ing a large amoun t Of calcium su lphat e in solution , and

further charged with lime from the neighbouring shore and reefs. First t he surface, exposed to t he t ide

,has i t s pores t o some extent fi l led up by sand particles. It further

hardens,and becomes thoroughly indurated with lime

,precipit at ed from t he wat er, which

dries on the surface aft er each ebb. Y et wit hal solu t ion is qu it e marked ; t he wavesstream Off by t ortuous channels worn in t he sand and nul lipore format ions between t he

harder corals,which

,if branched

,are left as an upstanding forest Of stems. B ehind t he

surface loose sand becomes more consolidat ed t han where t here is rain - w at er action alone.The solvent act ion wi t hin t he masses , save a lit t le due t o the rain , is inconsiderable, andthe whole differs very lit tle from the original reef

,except in i t s bet ter consolidat ion and

t hefi l ling in Of some of i t s int erspaces.

The conglomerate in M in ikoi is found everywhere at t he base of t he ou t er beach ,especially a t t he poin t s. Immense masses occur in t he beach on t he lagoon side of Kodipoint (Pl. I. fig. and continue in smal ler pinnacles along t he same shore , as far ast he rocky beach extends. The masses are much pitt ed and very rough on the surfaceand Of a slate - grey colour ; t hey consist of a consolidat ion of corals , nullipores, shel ls, et c.Between t ide- marks t he scour, sweeping t he sand and pebbles t o and fro, undermines andfinally precipitates the masses

,to wear t hemselves beds perhaps in t he rock beneat h , to

which t hey may become consolidated. I t fol lows t hen t hat no results can be Ob tained byt he examination Of t he smaller masses in respec t t o the posi t ion of t he corals, etc., ofwhich t hey are composed. The large lagoon masses near Kodi point are Of rough surfacewit h t heir corals freely exposed (Fig. Madrepor a is by far t he most abundant genus

1 Thesededuct ions arelargely based on my observa t ions and t he adjacent is lands. SeeRep or t B r i t. Asa , p. 400,on indubi tably ra ised rocks in theFijies , but morepar t ien 1900.

larly in thenorthern dis t rict s ofCey lon ,theJ afi

'

na Peninsula


represented,but Poci llopor a , Euphylli a , Pavoni a , Astraea , Ga lawea and Montipora were also

found. The modern facies or species Of M adrepor a grow from a cent ral mass wit h branchesradiat ing out over t heir upper surfaces. The central st ems grow s t raight upright , vert icallyindeed

,and are packed al l over with coral lit es

,possessing more or less raised calicles , which

F IG. 9. Conglomeratemass near W ir ingili a t low t ide, show ing the rugged appearance and undermin ing by t ida lact ion. (From a p hotograp h.)

give them a very rough appearance. The branches e x t end more and more horizontal ly fromt he cen t re ou t wards, but none grow downwards. The upper surfaces of the horizont albranches are similar t o t he cent ral stems, but their lower are smoot h

,coral lit es distant

from one anot her, calicl es wit h no raised rims. The basal mass,in which the branches

appear to be root ed , is formed by a fi l l ing up Of t he interspaces be t ween t he branchesby a ligh ter coral lum of - Open t exture , a kind of coenenchyma. Such be ing the mode ofgrowth Of M adrep or a i t fol low s t hat this genus forms an excel lent criterion by which toj udge whet her the conglomerate masses originally grew as such

,or were otherwise formed

,

as by a beach consolidation (PI. 11. fig.

On thepinna cles and masses thebr anches of thedifferent fa cies of Madrepora aremuchw orn but genera lly st i ll ent i re. Some of the colonies seem to be a li t tle t i lted, but most are

clear ly in the absolute p osi t i on in w hi ch they or igina lly grewl. Other genera of cora ls

amply confirm theevi denceof eleva t i on, w hi ch thi s fa ct imp lies , and imbedded Tridacna shells

w i th thei r open mouths in a hor i z onta l p osi t i on lend thei r supp or t.

It remains to consider the posi t ion in which the conglomerat e was originally formed.The fine- grained struct ure Of some parts shows t he presence Of nul lipores , but t he lat t erare never in sufficient quant i ty, al lowing amply for denudat ion , for the whole t o haveformed part Of t he surface Of a reef awash. At t he same time spreading nul lipores

,of

which t here are indubi t able traces, show t hat t he rock must have been formed in a reef,

freely ex posed t o the sea, and cannot have been built up in a lagoon. At M inikoi I was

1 I t occurred to metha t thespecific gravity of thebasa l Thebranches werebroken off, thebasal masses a loneweremass migh t be such that theM adrepora would nat ura l ly left . They lay in every conceivablepos it ion , moreindeed on

tend t o t akeup th is pos it ion. I hencehad a largenumber theupper surface,as w ithout i t s branches i t becamea lmost

of colonies (abou t 30) placed on thereef near Teveratu point. fla t.

36 J. STANLEY CARDINER.

never able to approach a typical pa rt Of the reef from seaw ardl,but in the Maldives at

about three fa thoms t he bot t om consist s Of great butt resses and masses wi t h branchingM adrep ora and Poci llop ora ,

massive corals here and there, and a considerable nulliporegrow t h between. These coral colonies are for the most part smal l , t he branches short andstout

,—t he usual

‘condit ion near t he surface, or in any strong current,—and somewhat re

sembling those of t he conglomerat e masses. Between the bu t t resses sand and a few bou ldersof coral

,broken off t he higher parts of the reef and rolled smoo t h

,are seen

, t hemselvesperhaps to be bui lt in t o the reef. “ Such boulders I have not found in t he conglomerate

,and

t heir absence would indicate t hat the lat t er was formed bot h deeper and before t heexist enceOf any reefflat t o t he at ol l. I ndeed the conglomera te w ould seem to have cons ti tuted p ar t

of the ori gi na l reef (w hi ch. by subsequent grow th fashioned the a toll of M i nikoi ) a t a t ime

antecedent to theforma ti on of any defini tereef-fla t , and a t a dep th of a t lea st threefa thoms.The conglomera te then may be considered to p rove conclus ively an eleva tion of thea toll. I ts

summi ts are 6 feet above the reef—fla t (the low - t ide level) and,

allow ing 18 feet for the

dep th a t w hi ch i t w as formed, there mus t have been an a ltera ti on of level of a t least

24 feet.Re t urning to the condition of the whole reef at t he t ime Of elevat ion

,i f there were

a broad reefflat , no addit ional upheaval t o t hat suggest ed above would be required t o

explain t he heigh t of the island in different positions. The boulders , which form t he rockyarea

,he more or less horizont al ly on one anot her

,and are much rain - worn. I f they had

been piled up by the waves , t hey wou ld show a definit e dip , but their position is morein accordance wi t h t heir being formed by the simple sinking Of the surface owing to t he

solvent action Of the rain , removing nul lipores, sand and the soft er part s. The heigh t oft he land - sections D and F is abou t 14 feet z

,and t he elevat ion based on ot her considera

tions 24 feet , giving a sinking Of 10 feet t o t he surface. Considering the heavy rainfall,

the spongy nat ure of a reef, and t he almost complete absence of evidence Of nulliporegrow t h in t he boulders, this erosion Of t he surface wou ld seem to me but reasonable.Allowing an average loss Of half an inch in a cen t ury, ample even if t he original reef werenot solid ,

only years since t he elevat ion wou ld be required. This period is perhapsmore t han sufficient for all t he changes which have t aken place in the at ol l since i t s elevation

,

but is not synchronous wi t h any known upheaval , such as migh t nat ural ly be ex pectedin t he H indust an continent

,if such a change Of level has t aken place , as I believe, more

or less a t the same t ime throughou t bo t h t heLaccadives and Maldives.

Let us proceed however t o an e x aminat ion Of t he sandy areas , so as t o see whetherthey lend any support to t he changes t hat have so far been deduced. I have alreadydescribed the surface charact ers of t his area wit h t heexcept ion Of a smal l patch at BoniKodi

,similar sand t o w hich indeed forms a bel t everywhere against t he rocky area. It

consist s of coarse part icles of all t he organic skeletons , which assist in forming a reef, w ithlarger rounded pieces Of coral. I t t hus differs en t ire ly from the fine sand Of t he lagoonshore elsewhere , resembling somewhat closely t he

.sand a t the base of t he sea - beach of t he

south half Of t he island. I invest igated i t s distribution by digging pits in sect ion across

1 For t hereef characters in different pos i t ions seeSec.III. found in sect ion G , as I cons ider tha t theremay havebeenS uch reef as ex ist s between Mou -Rambu and Kodi poin t s i s a long the ridgeherea piling up of boulders , even land, a t anot typica l. t imeanteceden t t o theelevat ion.

2 I havenot t aken the grea ter eleva t ion—19 feet—as

TH E ATOLL OF M INIKOI. 37

the island and by examining al l wel ls and na t ive excava t ions. Nort h Of the vi l lage i t

practical ly extends from t he old raised reef to t he lagoon , but southwards i t graduallydecreases in breadt h , giving place t o t he regu lar lagoon beach sand. Towards the rockyarea more coral fragment s Occur , but t he j unct ion is sharply defined , while on i t s lagoonside i t merges very gradually in t o the finer sand. The pa t ch a t Boni - Kodi forms a soft ,very friable sands t one on the surfa ce wi t h loose sand benea t h , apparen t ly t he crowbar final lyfet ching a hard rock. The amount of consol idation elsewhere is very varied , bu t a t t he

mean high - t ide level it commonly forms a loose sandstone, increasing gradually in hardnesst o t he low - t ide limit , below which soft sand succeeds and w at er is obtained , varying insweetness wit h t hedist ance from t he sea.

Reef-fla ts w i thout i slands genera lly slope tow ards their lagoons for some dis ta nce, the

hard rock gr adua lly giving place to a coar se sa nd, w hi ch has been swep t acr oss thefla t bythe w aves‘. The origin of thi s belt of coar se sand a t M inikoi mus t have been, I think, the

same, and i ts occur rencehencesuppor ts theview tha t therocky area a t one t imeformed p ar tof a regula r reef-fla t.

Remembering that t he remainder of t he sand is certainly Of lagoon origin,i t is Of

interest t o remark that in some of my pi t s in it I found alternations Of harder and soft ersandst one, or loose sand. I t was impracticable t o t rench across, but my pits were sufficient lynumerous to warrant me in the belief that t here is an irregular series Of lines of sandrock

,more or less paral le l t o the lagoon shore with areas Of loose sand between. I shal l

have occasion subsequently to point out t hat the formation of sand - rock is characteristic ofbeaches, which are washing away , or at rest. Hence i t may perhap s be deduced tha t there

havebeen per i ods a t M inikoi , w hen sand w as being a ct ively heaped up from the lagoon ,a lter

na ting w i th per i ods of rest,or even of w a shing aw ay.

Beach - sand - rock a t t he present day crops out nowhere on the lagoon shores Of t he island.A lit tle only is found at the west end and On t he outer beach to the sou t h. In KO - Varibay i t fringes the beach against t he reef, bu t i t is doub t ful whet her its const it uent s didnot rat her form part Of t he original raised reef. The same remark perhaps appl ies t o i t s

presence elsewhere , but t hepoint is immat erial , a s no deduct ions are based on i t s occurrence.

The island Of W iringili is formed almost entirely Of rock,much of i t s surface being

bare. It is sit uated at one Of the most exposed parts Of the atolls,for t he sout h - wes t

gales are always those of t he great es t force,and t he reef outside i t is narrow. The con

glomerate rock forms a broad belt Of masses and pinnacles a t the inner par t of t he reef,many Of t hem much undermined (fig. they reach a maximum height over the reeffla t Of 8 feet , and loose masses add an addi t ional 2 or 3 feet t o t he alti t ude of the land.On the lagoon side of t he island is a col lec t ion of smal l fragment s of rock and sand. Theerosion on t he seaward face is very marked

,but the rest Of t he island is pro t ec t ed by

the conglomerat e band.

B agandi is a bare islet Of conglomerat e close to t he edge of the ree f. It is coveredusual ly wit h loose bou lders

,thrown up by t he waves, but the sea occasiona l ly sw eeps i t

quite clean , generally carrying away a t the same t ime a wooden beacon,which crowns i t s

summi t . It is probable t hat in a few years no traceOf t he land will remain.

1 Submerged reefs on the other h and are genera lly indeed any t ran s it ion to a sandy area.

precipitous on all s ides , and do not hence show such or

CHAPTER IV.

THE ATOLL OF M INIKOI (Cont inued).

SECTION 3. THE ATOLL REEF.The encircling reef Of M inikoi in its different aspects may be considered as typical Of

the whole of t he Maldive and Laccadive at ol l - reefs. In a section t hree parts may bedistinguished—the reefflat , the rough or boulder zone, and t he inner slope. The two formerare to some extent sharply separat ed Off, as their basis is a solid flat Of rock

,while the

third may be ent irely formed Of loose sand. There is also the ou t er slope to the depthsOf t hesea, but I have already referred to t his in my general accoun t Of t heatoll (Sec.1 , p.I shal l

,however

,recur to i t s charact ers in any posit ion where it may appear different from

the previous descrip t ions. On account of the variat ion it wil l be most convenient to sket chin t he charact ers Of t he reef in several different situations in a t raverse of the wholecircumference Of t he atol l. In my account of t he Maldives I shal l as far as possible fordescript ion merely refer to t he reef in differen t si t uations a t M inikoi at ol l.

Kodi point is essentially a sharp project ion, t hebreadth being less than 100 yards froma dep t h of I t fathoms on either side. Owing t o its si t uat ion by t he lagoon passage t he

current s in and out Of which , mee t ing w it h those along the seaward reef, form a whirlround t he point , t he sea breaks heavi ly on either side almost independently of the quarterfrom which the wind may happen to be blowing. H ence the two sides

,experiencing almost

t he same condit ions, resemble one ano t her closely. The point of t he reef proj ects about1 cable (200 yards) from t he land , and in its centre lies a series of conglomerate masses,40 yards long, running paral lel to t he outer edge of t he reef on eit her side. Taking asect ion across t he cen t re Of t he point, one finds t w o or three li tt le terraces on the ou t eror east side , each perhaps 12 yards broad and rising by a 2- feet step at its ou t er edge.They are st udded with low pinnacles of the conglomerate , which also form a broken lineat t he outer edge of each. These masses increase in size and height inwards

,and areOf

smoo t h cont our, exhibit ing ex t ernal ly no trace Of their const itu t ion. The whole surfacebetween is covered wit h low green algae, and is completely desti t u t e Of sand and bouldersof any sort ; no corals grow ,

and only a few low incrusting nul lipores are found towardsthe seaward edge of the ou t er t errace. Furt her seaward the whole bottom could be seenat the backwash , preceding each breaker, to be Spread wi t h red

,whit e and pink Li tho

thamnion toget her w ith a few corals, principally smal l colonies of the most massive facies Of

Poci llopora and M adrepor a (Pl. II. fig.


F i g. 1 . Reef from Kodi poin t , M in iko i , look ing nor t h

F ig . 2. Cong lomera tema ss a t ba seof beac hTofa cep.39

TH E ATOLL or M INIKOI. 39

In the centre of the reef the rock masses are all much eroded underneath by t he

seas,which at high t ide, breaking on eit her side , meet one anot her in the middle of the

reef to rush along and escape a t its north point . This central pinnacled fla t is abou t30 yards broad , and in places a few coral masses lie, some perhaps thrown up by the sea ,

bu t mos t eroded out of the conglomerat e. Between the masses a few pits are found,from

which t he water does not escape , and in t hese flourish colonies of Por i tes arenosa andPsammocor a p li ca ta (or similar facies). The lagoon face against t he passage is similar t othe seaward , but t he t erraced arrangement is less not iceable. The pinnacles are fewer andsmal ler

,and the bread t h is only about 30 yards. The drop outside is more abrup t

,t he

reef ending in a cliff of 15 t o 3 fathoms against the passage. The surface is smoot h,bu t

various larger foliaceous and calcareous algae (Ha limeda , etc.) flourish ; neit her nu llipores , norcorals are found even at the extreme edge. Round the end of the point the condit ionsof the two sides merge int o one another, but t he t erraces are broader ; the outer is not

marked wi t h pinnacles,and slopes gradual ly to the depths be low.

A series of large table - t opped dead rocks are seen on t hebackwash of the breakers outside the cliff against the lagoon passage ; t hey ex tend up from the level of the cent ralconglomerat e masses in a line to t he point of the reef. Owing t o

"

t heheavy seas I neversucceeded in making good my footing on any of these, but t hey appeared t o be pi t ted andbare of al l organic growth except a few

,low

,green algae. They are apparently being eroded

underneath,subsequent ly to topple over and to be comple t ely removed in a similar manner

to t he conglomerate masses of t he shore] I cannot find any record of such bare massesin the numerous accounts of coral reefs. In t hose

,which I have seen , I can only compare

them to t he table - topped , raised reefs in some of the Lau islands of theFijies ; the erosionof the latter

,however

,is mainly between tidal l imits. They may be the t ables of much

taller mushrooms, t he stalks of which have given way , while t he tables have not yet been

cut down by the sea to t he regular level in this position. I am,however , r a ther inclined

to believe tha t they aredue to a broadening of thep assage, and tha t they show thema x imum

efiects w hi ch can bep roduced by thecombined current and breaker a ct i ons on thereef. A reefflat cons ists of bu t t resses and masses which by organic growth have become j oined t ogether.The water has got behind some of t hese

,and t he present t able - shape resul t s. The force

of the waves w ould cause a rapid denudat ion of t he consolidat ing sand , and t he const an tchange of the wat er

,due to the curren t s of the passage, an equally quick solution. The cut

rushing water,carrying a large amoun t of dirt of al l sorts in suspension , would prevent any

organic growth , so that the former actions wou ld be unimpeded.

Between Kodi and Mon - Rambu point s t he reef is very similar to t hat eas twards ofthe former. There is a st eep

,boulder - s t rew n beach

,and below this a series of terraces

usually t hree with a step of 1 foot to each—before t he regu lar ou t ward slope commences.The pinnacles of conglomerat e are sparser

,bu t usually a well - defined row marks t he outer

limit of al l except t he most seaward,which is never uncovered a t any tide. The

_ lat t erterrace is not wel l defined

,being a fla t for a cert ain dist ance and then t ailing off sea

ward into the outer slope. I t s inner part is completely covered with low,green algae ,

which to seaward give place to large nu llipore areas with occasional colonies of M adrepora ,

Poci tlopor a ,COelori a and Psammocora in the hol lows. The inner terraces are t o some ex t ent

covered wit h green a lgae,but neither nul lipores nor corals are found except very rarely in


shal low pi t s. Often ba re p a tches of rock occur , the p innacles a re cons tantly falling , and

there i s no doubt bu t tha t they are being r ap idly eroded to the level of the outer ter ra ce.

Off Mon - Rambu t here are four t erraces , t he most seaward with t he same characters asthe ou t er of t hose described above , but abou t 35 yards in breadt h. The second has on lya very few pinnacles

,most ly a t the outer edge. The t hird is studded with blocks

,t he

out er limit being marked by an almos t complete l ine, and the fourt h , or innermost , is arocky mass of conglomerat e w i th top awash at ordinary high tide , broken up by a numberof fissures

, t hrough which the water escapes after each successive wave.In Ko- Vari bay the inner t errace forms a ridge at t hebase of the beach

, 3 4k feet abovet he low - tide limit . I t is qu i t e smooth , and formed of a conglomerate of sand and coralfragments. Part s consist of the ordinary coral conglomerat e , bu t most is of beach format ion

(see page I t is being worn away on t he surface, eroded underneat h , and great slabsare const antly being broken off, t he ac t ion nevert heless being much slower than on t he

coral - rock. The reef ou t side this inner t errace varies in breadt h to 45 yards. Two terraces ,each 10 to 12 yards across , occur, t he inner smooth and bare of organic growt h with a wel lmarked line of pinnacles on i t s ou ter edge, t he seaward 1 foot lower, somewhat hollowedout in t he cen t re , edge mark ed only by an occasional pinnacle and t he surface coveredwith green algal growt hs w i t h nul lipores t o seaward. An outer fla t starts 1 foot lower ;i t varies up to 25 yards in bread t h , and slopes gradual ly to a dept h of about 5 feet ,where it gives place to a succession of coral and nul lipore masses. I t s surface is coveredwit h green algae

,nu llipores and corals of t he usual genera in addi t ion to massive and

branching facies of Por i tes.

From Ko- Vari bay westwards t he t errace formation disappears , giving place to a broadreef- fla t . Masses of conglomerate, however, crop out a t t hebase of the beach right round tothe end of t he island. Off Tevera t u point the beach is very st eep, —being formed of largeboulders from the conglomerat e

,—and merges below into the reef, which is about 80 yards

across. Three part s may be readily dist inguished ; an inner 20 yards broad , bare with a fewrock masses

, t he remains of the t errace formation ; a midd le pa rt, 40 yards across, nevercomplet ely uncovered a t low t ide

,t he t ypical reeffla t ; and an outer part , which slopes for

abou t 1 foot in a breadt h of 20 yards , consis t ing of great bu t tresses , perhaps divided across ,t he fissure zone. The reefflat is not iceable in t his posit ion for t he remarkable abundanceand luxuriance of cora l growth , as compared w it h t he same part to the sou t h and westof t he at ol l

,and general ly w i t h the Maldivan reefs. Three or four fac ies l of Madrep ora

abound , w it h low branches and spreading masses , in places covering over 20 per cent . oft he surface. Other corals spread over an addit ional 20 per cent., Coelor i a daeda lea ,

Poci llo

pora coesp i tosa , Lep tor i a tenui s. Orbicella ( t w o or t hree species), Pavonia repens. Psammocora

p li ca ta and Por i tes pa lma ta being t hemost numerous ? The fissure zone is simply a part

1 By t he term facies I imply s imply a modeof growth.I cannot use the term species in referring to M adrepora ;

i t i s too precise, as I am uncer t a in whether all, w h ileappearing dis t inc t , are not rea l ly the same Species. The

determina t ions of the Species , wh ile usua lly accura te, are

on ly in tended t o convey an idea of thefacies of thegenera

represen ted. Moremass ivespecies ofPoci llopora and Pori tes

werea lso obta ined in addit ion to specimens of Pr ionas traea ,H ydnop hom, G oni a s traea and Astr aea .

I am a t a loss for a completeex plana t ion of th is grea tlux ur ianceof cora l growth , unpara l leled i n my ex per ienceof similar reeffla t s , both for quan t i ty of cora l and var ietyof genera and species. Tevera tu poin t project s to thesou theast , eigh t point s from ei ther of theprevai l ing winds , t he

THE ATOLL OF M INIKOI. 4 1

of t he reefflat, which is broken into bu t t resses and masses by the channels along whicht he superincumbent water escapes. The fissures vary enormously

,some being only a few

inches across,ot hers many yards , serving perhaps as boat - channels

, t he waves surgingthrough and not piling up as breakers. The channels gradual ly slope from the reef to adepth of about 21} fathoms at t he ends of t he bu t t resses. The lat t er may be of almostany size

,and are often much in t ersect ed. The rol lers break just outside their seaward

ends,which in any one position lie approximately in line with one anot her. Off Tevera tu

t he buttresses are smooth on t he top , covered with seaweeds , with al l the species of coralsof the true reef- fl at , t ogether wit h many stunted colonies of some massive facies of Poci llo

p ora. To seaward they end more or less precipitously , but towards t hefissures oft en overhang from 1 to 2 feet ; their edges and sides—e x cept where they are overhung—are completely covered with encrust ing nul lipores. Some of the channels extend right int o t he trueflat , bu t none are large enough for boats. They have no coral growth , but boulders of recentcoral origin may lie in t hem , general ly more or less bedded in sand , perhaps being consolidat ed t o t he rock by nul lipores.

The reef off the lighthouse and t he west ern part s of the island exhibit s t he samedivisions as at Tevera t u ,

but t he true reef- fia t is abou t twice as broad. The beach isformed of small boulders and sand , t ailing off in t o a rough area

,20 yards broad , which

represents t he inner zone of Teverat u. I t has t he same characters as t he latter but lessin t ensified

,and furt her is s t rewn wit h coral masses , some thrown up by the waves, bu t

the majori ty washed down from t he land behind. I cou ld find no trace of any consolidationgoing on at t he present day, the loose masses usually lying bedded in a lit tle sand , whichmay have a few crustaceans and worms

,perhaps al so a holo t hurian or two and some

actinians. The under sides of t he blocks , where exposed freely t o the water, are bare , orcovered in pat ches wi t h t hin sponge and Tunicate colonies. The broad flat is almos t devoidof life of any sort ; its surfa ce has a somewhat slimy appearance from the mud and dirt.Such algae as grow on i t are smal l and fi lament ous

,si lt ed at their roots with mud . A few

smal l pits are found,having perchance a bou lder or two bedded in sand

,t hewhole of the

rest of the hol low oblit erat ed by Halimeda,or some similar alga with calcareous leaves.

The fissure zone is t he same bread t h as off Teverat u,but it is higher

,and resembles

more the same zone,described by me at Funafut i and Rot uma l. I t s surface is pitted.and

almost completely covered wi t h low green algae and nullipores. In larger pits corals maygrow, but the colonies are of smal l size ", a species resembling Poci llopora coesp i tosa beingthe most common ; an incrusting M ont ipora alone at t ains any magni tude, although most ofthe genera found a t Tevera tu may be represented. The seaward edge of the zone forms

ordinary current s caused by wh ich would bebound to sweepround i t . Probably theseas a lways break on i t heavi ly ,keeping thewater of thereef- flat cool , and prac t ica l ly a t t he

lowes t t ides preven t ing any dry ing up of thecora l polyps.Thewholereef, too, having a rocky shore, i s s ingular ly freefrommud and dir t of all sor t s. On thewholeI am inclinedto th ink tha t th is lux urianceon ly dates back for t wo or three

seasons, wh ich havebeen pecul iar ly favourable. Thesmal ls i zeof most of the cora l growth s lends th is View some

support.1 TheCoral Reefs of Funa fut i , Rot uma and F ij i , et c.G.

Proc.Camb.Phi l.Soc., vol.1x ., pp.417 - 503.

2 Th is i s a mos t remarkable fea t urea long thewholereefin th is pos it ion , few of thecolonies being morethan a few

inches across. At thecommencemen t of t hemuch delayedsouth -wes t monsoon in Augus t , 1899, thebeach w as s t rewnw ith sma l l cora l fragment s , and thereseemed to becons iderabledea th among thereef cora ls , owing to dirt and mud.

I am inclined to th ink that mos t of t he branch ing cora lcolon ies had been formed s incethela st sou th -wes t monsoon

(in t h is case10months), and th a t few or nonesurvivei t.


a marked line,and the fissures are small , many closed over by nu llipores , which may

perhaps leave smal l blowholes.The reef between t he last two positions described gradual ly varies from the one to

t he ot her. In one place for abou t 200 yards the whole of t he seaward face of the buttresses and for some distance up the channels is singular in being absolu tely covered overby a yel low zoan t hid actinian. Off W iringili and B agandi islands the reef is nearly thesame as off Tevera t u point

,but t he conglomerat e zone is broader

,more marked , and very

rugged (Fig. Between W iringili and M in ikoi the reef out side t he boulder zone is

F IG. 10. Seaward beach of W ir ingili with masses of conglomerate. (From a.p hotograp h.)

intermediat e between the reef off the end of t he large island and the reef of t he wholeof the west and north sides of t he at oll. The latter does not differ material ly in differentparts

,and t he descrip t ion in one posit ion wi ll serve for the whole.

In a cross section between W i ringili and B agandi three zones are wel l defined, t he

fissure,reefflat

,and boulder. The fissure zone does not very mat erial ly differ from t hat off

t he end of Minikoi island , but its edge against t he sea is more irregular, some massesdropping wi t h a cliff t o t w o or t hree fathoms , others t ailing off gradual ly. I t varies inbreadt h up t o 30 yards , and slopes about two feet i n this dist ance from i t s crest ; t he

lat ter is a sligh t ly higher part j ust e x posed at low springs, six or eigh t yards across. Furtherou t side the slope quickly at t a ins a dept h of four or five fat homs

,slow ly increasing for some

dist ance. The bo t t om is fairly smoo t h,covered wit h spreading nullipores and s t udded with

round masses of the same and coral colonies , t he former always wit h a cora l core. Thereis no appearance of the great masses found off t he reef t o t he sou t h of M inikoi island.The seas break just outside t he edge of t he zone, which except the crest is swept by t he

waves in al l weat her. The fissures seldom have overhanging sides,or any accumulat ion of

coral blocks or sand in t heir channels. They generally s t art in t he cres t and slope t o

about 295 fat homs out side ; few reach t hrough to t he reefflat,from which at low t ide t he

water cannot read i ly escape seaward. The water a t low t ide rushes up t he fissures anda cert ain amount wel ls over on to t he reeffla t

,t hough mos t

,being t hrown back by t he

crest , escapes wit h great force down t he channels. The bu t t resses are pi t ted on t he surfaceinto smal l rounded hol lows

,which are oft en fi l led wit h algae

,al lied t o H a limeda

,or with

smal l coral colonies,mos t ly branching facies of t he usual genera. G eneral ly, however, t he

hollows are empty , t he rock beneat h merely covered w it h a nu llipore, or a fla t green algaof some sort. If t he surface be broken

,a nucleus

,usual ly of coral

,is reached after several

inches of a ligh t concret ion of worm t ubes and Li thothamnion. The walls and floors of the


fissures are usually smooth , and encrus ted completely wit h nul lipores. Free animal life ofany sort is rare; a few t hick - spined Echinoids and large crabs wedge themselves int o thehol lows ; bri l lian t ly coloured fish wash to and fro wi t h the w aves in t he fissures ; betweenthe branches of corals and weed "a few Alphaeids find a refuge with an occasional Cypridand Stromb. The crest of the zone is not always present in such a marked manner

,though

an inner part , densely covered wi t h green algae , in which the fissures most ly end,can

alw ays be dist inguished. I t s surface has t he regular worm tube and nullipore structure,only

in great er thickness ; corals are rare , a few mammillated nul lipores being t he only calcareousorganisms.

The reeffla t is about 50 yards in breadth , presenting almost the same charact ers asin t he last position described. The greater part of i ts surface , being sit uated abou t 6 inchesbelow t he crest of the outer zone, is a t any stat e of the tide covered with water

,in which

there mus t be some change through the fissures. I t s surface is absolu t ely free from loosecoral boulders

,and is for t he great er part bare save for low algal growt hs. It is clean

,with

no dirty appearance from mud or sand , and , i f broken into, is seen to be formed of hard coralrock. Small incrustations of nul lipores are found , but t hey are never of any real importance

,nor can t hey in any way affect t he general heigh t of the whole. The same genera

of corals grow as are found on t he fissure zone ; spreading forms of M ont ipor a alone at t ainany size

,t heir colonies sometimes covering 2 or 3 square yards. The most abundan t genus

is Poci llopor a ,of which two forms , finely and coarsely branching

,are numerous ; Coelor ia

daeda lea also is of considerable importance. Corals at t ract the eye,and hence tend t o

assume an undue signification. I t is very unusual for them to cover more than 10 per cent.of t he surface of any reef, sufficien t for wastage , bu t not enough t o alt er t he general level.Towards the exterior masses of hard rock , 2 to 3 feet high , are found irregularly abou tevery yards. They are pit ted on the surface and rott en

,perhaps the outside a mass of

worm tubes,surrounding a solid rock core. Although I passed along t his reef in the be

ginning of Sept ember,1899,

aft er very heavy south - west gales,I failed to find any large

masses t hrown up by t he seas (negro—heads) , nor did I ever find any rocks on the flat inprocess of cemen t ation on to t he reef. B locks added thus usual ly can be easi ly split off

wit h chisels along t he line of cement ation , but I failed t o break any of t hese. I am hence

driven to the conclus ion tha t thesemasses are truep inna cles , tha t they once absolutely were

pa r t of the reef i tself and were left , whi le the res t w as comp letely w a shed aw ay. Of freeanimal life

,ot her t han t hat found on the fissure zone

,Holothurians alone are prominent .

Of these two zones as far as the nort h passage in t o the atol l l i t t le further need besaid. They have nearly t he same characteristics

, but the fissure zone increasingly becomesmore broken and less defined a t its outer edge, which tai ls off very general ly by similarmasses t o t he seaward dept hs. The crest is in many parts absent

,and t he two zones merge

into one'

anot her ; pinnacles are scarcer. In three places between B agandi and t he passagethe reef is lower

,forming boat channels, which are used“ above half- tide in the nort h - east

monsoon. These passages are merely ex t ra large fissures,which run furt her in t o the reef

fla t ; t hey are kept open mainly by the scour of the tide,preven t ing coral and other

calcareous growths.

The boulder zone (Fig. 1 1) is an area of rough blocks of rock , separat ing the reef—flatfrom the lagoon. It is abou t 30 yards broad , moderately sharply defined on bot h sides ;

6—2


some of i t s masses attain a heigh t of 3 feet above the reef- fia t,but the general level is

1 to 2 feet less. What appears t o be t he inner part of t he reefflat forms t he foundat ion ;i t is to some ex t ent worn into holes , in which the larger loose masses lie, imbedded in sand.The three i slands merge a t their ends in t o t his zone

,masses of the conglomerate ex t ending

ou t from each for 100 yards or more along i t . It has a twofold origin,consisting of

boulders cast into this posit ion from t he reef and lagoon,and also of masses of the con

glomerate, some loose and some att ached to t he solid platform below. Relatively lit tle ofthe whole , however, consist s of masses attached to the reef, t he great er part being formedof loose blocks of t he old coral - rock. Towards t he ou t er side a certain number of coralblocks are thrown up from seaward , and on the lagoon side also a few may be washed

FIG. 11. Boulder zoneand reef - da t bet ween W iringili and B agandi. (From a photograph.)

up. The sand beneath is coarse , consisting of fragments of the conglomerat e and thewashings of t he reef. The boulders are singularly free from al l bori ng and o t her destruct iveorganisms ; t heir exposed surfaces are absolutely bare of any animal or plant life. Littlecan grow on t he basal reef on accoun t of the sand , and indeed no fixed organisms of anysort are found save on t he undersides of boulders , which may be almost covered wi t h t hinsponge, Tunicat e or Polyzoa colonies in numerous small incrust at ions wit h possibly a litt leHydroid and Actinian growth as wel l. B ril lian t ly coloured Turbellaria and similar lookingNudibranchiat a browse on t hese. Crabs of t hin compressed genera and Ophiurids hurry toescape on every side ; numerous Macrurans dart for shelter. Under the boulders a few sharpspined Echinids

,t hin—skinned Holot hurians and large Cyprids l ie free on t he surface

,and

in t he sand may be found an occasional Polychaet and Sipunculid. In places the boulderzonemay be lower, where t he wat er of t he rising t ide finds i t s earliest access t o t he lagoon ,and i t s surface pract ical ly smoot h and bare.

While t he bou lder zone has t heabove enumerated characters as far as B agandi , i t ceasesfurt her nort h t o be nearly so wel l defined. Indeed it is merely the inner part of thereef- fla t

,sparsely st rewn with boulders w i t h a few masses of the conglomerat e here and

t here. The lat ter do not exist opposi t e the boat channels,and are most numerous in ter

mediat e to these. Close t o t he nort h passage there are a large series of such masses,

many of great size,forming almost a bare conglomerat e flat

,covered with abou t a foot of

wat er at high t ide.



mass of rock and not a sand format ion. Its surface is muddy, slimy and worn into holesmore or less fi l led with sand ; i t is bare and smoo t h for t he great er part , and bouldersare pract ically non - exist ent. The whole area is singularly devoid of animal or plan t l ife

,

sessil e or free. A few species of coral grow at t he very edge, but on the surface I onlyfound a few small colonies of Pori tes a renosa ; delicat ely branching nul lipores may exist andalso H alimeda , but weed is general ly scarce.

The sand - flat at t he south of the main island extends up round t he reef t o t hewestand nort h

,covering about half the tot al area of t he lagoon. The whole

,except a smal l part

near t he l igh t house , is completely covered even a t low - tide springs , t hegreat er part having 2t o 3 feet of wat er. Around t he sand - flat , wherever t he reef exist s , an ou ter zone, 100 t o 1 50

yards broad,may be separated , characterised by i t s abundant growt h of corals , the rest of t he

sand being absolu t ely ba re. The slope from t he reef to t he general level of the fla t occursen t ire ly in this area. The principal coral is a species of M adrep ora wit h massive branches ,which is everywhere abundan t in this si t uation. Ot her more del icate Madrepores occur aswel l

,most ly in isolat ed colonies , but a st ag - horn species wit h branches up t o 1 inch in

diameter forms in places close t o the bou lder zone regular groves. Por i tes arenosa is foundeverywhere

,in masses up t o 20 fee t across, dead in the centre, but living at the edge,

and places where the curren t s are not great ly felt are often almost complet ely coveredby i t ; t he masses are always free , t heir st alks having been dest royed , and a re verygenerally somewhat hollowed out in t he centre. Pavoni a repens crops up anywhere , whilethe species of Poci llopor a ,

of which t here appear t o be several , are found only opposi t ebreaks in the boulder zone. Local ly e x t remely common , especially near the west ern andnort hern reefs

,are great masses of M i llepora ,

Heliop ora and Psammocora , while coloniesof Goni as traea , Astraea , Orbi cella , Pr i ona straea , S ideras traea , Mus sa and Euphyllia are oft enmet with

,frequently growing in the hol low masses of Por i tes arenosa. On t he under

surfaces of any of t he above Agar i cia may grow , and various species of Fungi a are veryabundant on the bare sand. H a limeda abounds, growing among t he basal branches ofM adrep ora ,

but o t her algae can scarcely be said to occur. A mere glance at t he abovelist of the more importan t genera serves to show the richness of the area, and the boldgrowt h of the colonies themselves proves t hat t he posit ion on t he whole is one eminen t lyfavourable to them. The part s poorest in corals lie immediately behind the bou lder zone ,where i t is especially high or low,

t he st i l lness of t he water and the t oo rapid currentsbo t h being unfavourable. Surveyed from a boat at high t ide the presence of breaks inthe bou lder zone can be assuredly told by t he whi t e , sandy bot tom cont inuing up to it.Over the richer areas at t he sides of these, heavy boo t s and putt ies are required , as onelit eral ly breaks one ’s way t hrough groves of coral. The skeletons of l iving cora ls (i .e. coveredby the l iving polyps) from t he area are singu larly free from destruc t ive organisms of al lsorts

,no Polychaet s nor Sipunculids ; even sponges, which pervade t he hard parts of most

or al l reef corals,do not exercise much sway. The grow th is hence probably ex t remely

rapid, a nd yet t he decay must be as quick. The area is cert ainly not increasing in height.Dead corals are found bored through and t hrough ; t hey soon rot away

,and break down

in t o sand , or aredissolved.

The reef of t he boulder zone ' can at i t s edge be very dist inc t ly traced under t he sand,w hich varies between t hat of t he former zone and t he fine mud of the cen t re of the at ol l.

THE ATOLL OF M INIK OI. 47

Indeed w it h a short crowbar t he presence of such a reef is indicat ed oft en for 30 t o 40

yards,and i t would appear t o underlie a not inconsiderable part of t he sand - flat . At t ached

organisms,ot her t han t hose which have been already mentioned , are rare in t his outer

area,but nearly al l t he free forms of t he boulder zone wi t h t he except ion of Ophiurids

are found in as great , or even great er abundance. The black beche - de - mer is ex t remelynumerous , and myriads of smal l fish dart between t he branches of the corals. Annel ids andSipuncu lids are rare

,bu t generally free li fe on the coral reefs of t he deep sea is nowhere

as abundan t nor varied as on tropical con t in ental coast s , or on even moderately rich groundof t heMedi t erranean and English seas.

The rest of t he sand - fla t is nearly bare of al l organic growth. Opposit e Mini koi is landtwo smal l green algae are found in patches, and here and t here clumps of a finely branching nu llipore grow

,nearly imbedded in sand. Surface- living Holothurians are very common

in places,and a smal l species of P tychodera occurs local ly under t he weed in great abnud

ance. Living in the sand are large numbers of a whit e Holot hurian , a large Sipunculidand t w o Synap ta , besides vast numbers of smal ler species of the same groups. In addit ionheaps of sand , 4 to 5 inches high

,in a calm sea stud t he whole, t he cast ings of an

immense Enteropneust , t heend of t he body of which is oft en 1% inches in diameterwhen dist ended wit h sand.

The sand ends about 100 yards outside the line,marked 2 fathoms in t he chart,

opposi te'

and to the nort h of M inikoi island , the int ermediate area genera l ly being coveredw i t h coral shoals , in tersected by channels wit h abou t 2 fathoms of wat er. These shoalsare usual ly flat on the top , bare or hollow ed out in the centre

,with perpendicular

,or

overhanging sides—especially towards the deeper part of the lagoon—covered wit h coralgrowth. On their summits the chief coral is Por i tes a renosa ,

but al l t he previouslymentioned lagoon genera are represen t ed ; massive Ast raeids tend to be t hedominant formson the sides. All are much infested wi t h boring organisms

,Li thodomus

, Cirripedes and aG ast ropod as w el l as t he usual Sipuncu lids and Polychaet s. The patches decrease in sizetowards the sand area, on which t hey seem on t he whole to be encroaching. To t he sou t hwest and nor t h of the a t ol l coral pat ches may occur near t he “

edge of t he sand - flat,but

th ere is no defini t e broad line. This is especial ly the case by t he boat channels andnorthern passage. By t he side of the lat ter I found what appeared t o be the remains ofa pat ch , w hich for some cause or ot her had been kil led t he previous year. I t was a roundmass

,st anding in 3 fat homs wit h 1 foot of wat er covering i t at low tide

,sides precipi t ous ,

diameter abou t 18 yards. I t s surface was rott en,i t s sides ex t remely sparsely covered wit h

l iving corals—the great er part dead at t he base—with much dead and rott ing coral. Fourother pat ches seemed t o be in t he same condition

,but they had a cert ain amount of

weed,and would appear t o have been kil led some t ime previously. Several o t her reefs in

t he vicini ty had also been affec t ed , but none t o t he same extent ; some , however, had hadin places quite sensible slips a t t he sides. Free animal l ife on al l t he shoals

,i t may be

noted,is ex t remely scarce.

The slope from 2 or 25 fat homs, the general depth of the edge of t he surroundingflat and reef, t o t he deepes t waters of the lagoon is gradual and regular. Coral pat chesoccur everywhere but much more sparingly in the deeper waters and t owards the nort h ;indeed t he cent re of the lagoon is fairly clear. The patches al l reach wit hin a couple offee t of t he surface. Wit h the lead I fai led t o find any incipien t pat ches, and I lined


the whole lagoon wit h dredgings wit hou t finding any trace of such. In its sou t hern halft his deeper area is covered wit h extremely fine sand wit h delicate shells and Halimedaleaflet s. This gradually gives place nort hwards to an area Of broken coral fragment s in

process of decay,s t udded perhaps wit h smal l colonies of Polyzoa.“ Towards the passage

Polytrema and nul lipores begin t o appear, and the bo t tom consist s largely of their roundedmasses

, 1 or 2 inches in diamet er, each on a coral core. The dept hs in the Admiral tyChart are t o al l intent s accurat e , but t he 2 fat hom area is great er, as also is the 5 fat hom.The lat t er indeed occupies t he whole centre, and in t he middle of it is an elongat ed pitwi t h 7 t o 9 fat homs. Dredging in t his part of t he lagoon is ex tremely unproduct ive.NO fixed organisms ot her than t hose already ment ioned Were ob tained. For the rest 2

G ast ropods, 1 Lamel libranch , an Asymmet ron , a Carib and a few sand - loving Macrura alone

were obt ained in any considerable numbers.

SECTION 5. THE H ISTORY OF M INIKOI As AN ATOLL.In t he foregoing I have been at some pains t o show that an elevation of a t least

24 fee t mus t be al lowed for t he whole at ol l. I do not myself consider t hat t he changeOf leve l has been much great er t han t his. Therein I am supported by the general appearanceof t he corals in t he conglomerate, which indicat es a format ion in shal low water. The pointmay seem a smal l one

,but i t gains in import ance from t he wide distribu t ion of absolutely

similar conglomerate on the at ol ls bot h of t he Indian and Pacific Oceans. Yet I cannotmaintain t hat an ocean reef a t a depth of 10 or even 1 5 fat homs

,swept

,as is probable

,

by a st rong current,would not i f e levat ed t o the same heigh t have in its conglomerate a

similar st ruct ure. The sea has now been quite sufi c ient ly explored in the t ropics to showthat relat ively few reefs exist of such a charact er, most having at least grow n int o trueat oll reefs awash. The reef, t oo

, would have had to have been caugh t by t he elevationa t j ust the righ t period, for w it h pure oceanic condit ions—t hemos t favourable, if judgementmay be based on t he more perfect forms of at ol ls , afar from land

,—i t could only t ake a

relat ively very short t ime for the reef to rise from 10, or even 20 or 30 fat homs t o t he

surface. The rocky area of the land l ies obviously on reef. Remembering bot h t his factand t he regulari ty Of the exis t ing reef, I may poin t out t hat such deep shoals have not

usual ly perfect rims , but merely here and there peaks and banks approaching t he surface.I t appear s to me then a fa i r a ssump t ion tha t the atoll ex isted as such when the change

of level took pla ce. In further support of my content ion t hat t he elevation has been butsmall

,I would point out that t he conglomerat e appears to have been of very recent forma

t ion , a charac t er fu l ly supported by t he same rock in t hewhole of t heMaldive Archipelago.Coral - rock, when elevat ed in any great t hickness

,soon loses i t s dist inct

,organic st ruct ure

,

assuming a hard,uniform

,crys t al line charact er by rain - wat er solut ion and t he reprec ipi t a

tion of i t s line. This may be seen in t he Fijies, Tonga, Pelews,Solomons

, Christmasisland and West Indies

,indeed wherever reefs of any considerable t hickness have been

elevated above t he waves. The island Of Vit i Levu , Fij i , is perhaps an except ion,but the

condit ions are peculiar here, in t hat there must have been several elevat ions and suhsidenoss, in which t he coral - rock was covered wi t h t he pract ically impervious so- called

Of course t he argument depends on t he assumpt ion wit h which I st art ed,

1 Vi deProc. Camb. Ph i l. S oc., vol. I x ., p.453.


but with each additional inch eroded the condition of t he rock shou ld differ furt her fromthat Of the modern reef

,while there is actual ly little or no change.

It remains then to consider the form of the atol l , when first elevat ed,as far as may

be deduced from the evidence before us. I have already point ed out that gain and lossare bot h taking place in respect to the land , and t he distribu t ion Of these t w o fact ors.F i rst i t i s obvious tha t the a toll has been sta tionary for a considerable per iod of t ime onaccount of the broad reefflat round most of i ts circumference , together wit h the lagoonreef off M inikoi island and the broad sand - fla t everywhere. Fur ther there are t he sameactions of gain and loss in respect t o the reef. Wit hout entering into detai l no one doubts

,

so far as I am aware, t hat a reef in a st ationary area, ex posed to the sea, would gradual lygrow seawards. The bu t t ressed edge of t he reef Of our atol l

,al l covered with calcareous

organisms , t he masses growing on t o i t,and the further masses seaward

,covered wi t h

nul lipores and corals , give indubit able presumption of gain from t he sea. On genera l

gr ounds and from ex per ience I consi der tha t the sea gains on the reef in lagoons , w here

the ci rcula t ion of w a ter i s“

cons iderable. Specifically I may a t M inikoi poin t t o the deadreefs near the nort h passage

,and to t he similar charact er of the whole surface and edge

of the reef, extending along from the vi llage t o t he nort h point l.

Seeing t he changes taking place I summoned al l t he Old men of M inikoi to my aid ,and by t heir assistance

.

I am enabled to recons t ruct t o some extent t he appearance of theat ol l at t he commencement of the las t cen t ury. First there has been l itt le apparent change ,eit her in the seaw ard part of t he reef

,or in t he ou t er beach , save only t hat t he outer

paved path was a t t hat t ime perfect from t he vi l lage northwards and separat ed by shrubsfrom the sea. The rocks at Kodi poin t further were crowned wi t h coconu t t rees, andj oined to the main island. B el t s of coconu t trees existed righ t up from the v i l lage t o

t he nort h Of the island along t he lagoon shore , while now only a few are found a t B oniKodi. Indeed the nat ives add an average breadth from memory (60 years) of 30 to 40yards, an amoun t which seems incredible in a closed lagoon , but is fully supported by t hestrew n coconut and o ther t rees on t his beach as wel l as the house remains a t Bali - miHuugi. The ac t ion cont inually goes on , but hurr icanes are especial ly dest ructive, causingslips everywhere along both lagoon and outer beaches. Against this must be recorded anextensive gain—60 t o 70 yards it is s t at ed—between the vil lage and Moli - Mati point, butTundu point has retreat ed almost a like distance

,t he intermediat e area varying. An island ,

Tori - Gandu , on the reef to t he wes t Of t he nort h passage existed wi t h 4 or 5 coconuttrees

,while B agand i at th e same t ime had as many as W iringili now

,i .e. about 30, but

the latter island is on ly S lightly smaller. Lastly i t was st ated t hat t he nort h passage hasbroadened great ly , but Views as to its deepening

,though held by the majority, produced

an acrimonious discussion,from which and on ot her grounds I am not inclined t o credi t any

considerable difference.

The presence of conglomerat e masses I can only regard as indicat ing t he existenceof former land in any posit ion

,where t hey now occur. The land then must have a t one

1 My soundings , as compared with the chart , gave misca lcula t ion in thestateOf thet ide, or in thepos it ions Ofevidenceof such incons iderablechangein thesamedirect ion thevar ious poin t s taken for fi x a t ion ,

migh t ex pla in thediffertha t I do not care to lay any emphasis on them. A sma l l ences of thet wo set s of soundings.G.

5 0 J. STANLEY GARDINER.

time ext ended round the whole atol l w it h only a single break perhaps to the north wi t hlower parts here and there , W here boat chann els across the reef now ex ist. The seawardshore mus t a t t hat t ime have been relatively steep, like t hat between Kodi and MouRambu poin t s at the present day but wi t hou t terraces. The Old reef must have hadextensive sand - flats

,extending out from it into the lagoon on every side

,and the lagoon

on change of level,if it existed a t al l

,must have been at most a mere small

,central

pool wit h a slight ebb and flow Of the tide through'

a single inconsiderable channel to thenort h. Indeed save for t his smal l channel M ini koi cannot at one t ime have been fardifferent from t he low coral islands such as Oa i t upu (E l lice G roup), Mulaku (Maldives) andTromelin, which exist in considerable numbers in bot h the Pacific and Indian Oceans.Supposing t he land in t he present atoll to be entirely swept away, the condition at thepresent day cannot be far different from t hat of t he atol l before the change of level ,allowing for i t s then smaller size.

(To becontinued.)

H Y M E NOPTE RA.

B Y P. CAMERON.

THE Hymenoptera taken by Mr Stanley G ardiner in t he Maldives and in the MinikoiAtol l number 25 , of which t wo are parasit ic (I chneumonidae), the others belonging to theAcu leate section of the order. The number of species captured on the Maldives is 20; i nM in ikoi 10; 5 species are common to both groups. The known species are Indian formsof wide distribution in the Oriental Zoological region ; and some of them , e.g. Cera tina

vir idi ssini a ,Xylocopa

'

aestuans, Poli stes s tigma , and P.hebraeus , extend also into Africa andthe Malay Archipelago. All the species, new and Old, belong to genera of universal dis t ribu t ion in temperate and tropical coun t ries. The genus Megachi le, for example, ranges fromthe borea l dist ricts Of Europe and America t o Australia

,and few of the Pacific Islands

are wi t hout a represent ative of i t.

The M inikoi species were captured from June to September during the south - westmonsoon

,which is usual ly the wet monsoon ; but in 1899 during the period of Mr G ardiner

’svisit the weather was abnormal ly dry, the rain not commencing unti l the second week inAugust. The Maldive species were t aken in November and December

,which were very

dry months ; but in October heavy rain fel l. Hulule Island was visited in January andFebruary , 1900.

HYMENOPTERA PARASITICA.

Fam. I c h neumon i dae.

1. Z anthopimp la appendi cula ta ,sp. nov.

Long. 1 1 ; terebra 4 mm.

A species close ly al lied to the doubtfu l Z .puncta ta Fab. and which is certainly differentfrom Z . puncta ta Krieger.

Hab.M inikoi , Laccadives.

Scape of antennae pale yel low ; the flagel lum brownish. Head yellow , except the ocel larregion

,where t here is a mark longer t han broad

,which extends to the end of the vert ex

and is rounded in front. The face is strongly and closely, the clypeus less strongly,punctured. The head is not much developed behind the eyes , and is roundly , obliquelynarrowed there. On the base of the mesonotum are three black marks touching each other ;

7—2

5 2 P. CAM ERON.t he central is about as wide as long, t ransverse at the base and apex

,and w ith t he sides

bu lging out roundly ; t he lateral marks are larger ; their base is st raight and slightlyoblique; their out er side is rounded outwardly ; on t he inner t hey are obliquely narrowed.On t he base Of t he median segment are t w o black , ovoid marks, placed t ransversely, t he

broad end on t he outer side ; t he basal area is wider t han long ; i t becomes gradual ly , butnot great ly

,widened t owards t he apex , which is t ransverse; t he outer basal areae are wider

t han long ; the area next to i t , in front , is triangular, Oblique,and longer t han t he widt h

of t he base , where t he bounding kee l bulges out. Pleurae smoot h and shining. Wingsclear hyaline ; the nervures and stigma are black ; t he areolet is dist inct ly appendiculat ed.Legs coloured like the body ; the hinder tibiae are black a t t he base. On the abdomenare twelve black marks ; the pair on the petiole are broader than long, and have a short ,sharp proj ect ion on t he inner side at the base ; the second pair are smal l ; the t hird aret he largest ; are about as wide as long and rounded on the inner side ; t he fourt h pair aresmaller and somewhat similar in shape ; the fi fth distinctly wider than long ; the six th aresti l l wider

,but not so long ; t here are no marks on t he six t h and apical segments. In the

3 t here are no marks on t he second segment ; that on the fourt h is very small ; inthe middle of t he last segment are two curved obl ique furrows.

2. Eni cosp i lus reti cula tus, sp. nov.

Lu t eo ; abdominis apice nigro ; met anot o ret iculat o ; alis hyalinis , st igmat-e fusco. 2.

Long. 14 mm.

H ab. H ulu le,Male At oll , Maldives.

Antennae luteous,darker towards t he apex. The sides of t he face and the clypeus

have a yel lowish t inge. The face is finely wrink led in the middle and punctured on eit herside ; the clypeus is smoot h and indist inctly punct ured at t he base ; t he labrum wants theyel low t int of t he clypeus ; t he vertex and the eye orbit s have a dist inct yel lowish tint .Mandibular teet h black. Mesonotum fuscous

,except at the base and sides. The scutellar

keels are stou t ; the apex is closely , longit udinal ly s t riated. The base of t he median segmen t is smoot h

,wi t h t he sides irregularly st riat ed ; the rest of i t is irregularly, dist inct ly

re t iculated ; on t he apex t he ret icu lat ions are oblique and less dist inct in t hemiddle below.

Pro and meso - pleurae closely punctured , the middle of the pro and t he lower part of t hemeso closely st riated. The met apleurae above shagreened ; t he lower apical part bears somest riae. Wings clear hyaline ; t he stigma and nervures fuscous ; t he basal abscissa of t he

cubit us is t hickened,t he basal t w o- t hirds being t hicker t han t he apical part ; t here are

t w o dist inct horny point s ; the lower one is di lated and rounded behind ; i t s upper part formsa somewhat t riangular shape ; the apical point is somewhat oval in shape. Legs colouredlike t hebody ; the binder tarsi are Short ly but dis t inct ly longer than t he tibiae ; t he ant eriortwice t heir lengt h.

The t ransverse median nervure is received short ly behind t he t ransverse basal ; the re

current nervure is bullat ed on t he t op ; the disco - cubital nervure more widely in frontof i t .


5 4 P. CAM ERON.

being almost in a line w it h t he lower, and of‘ almost equal l ength t o it : in idr ieus the

lower is longer than the upper, which is on a different angle with it, i t being only slightlyoblique.

The lower part of t he head , inside and out, is t hickly covered wit h si lvery pubescence ;t he clypeus is stou t ly keeled. The scape of t he antennae

, t he mandibles, excep t at theapex, an interrupted line on t he pronot um , the sides of the scu t el lum ,

t he scutel lar keels,

t he tubercles,a mark on the upper side Of the mesopleurae, between the two furrows , a

longish line on t he sides of t he t hird segmqnt at the base, an oval mark on i t s sides att he apex

,and a short er line on the sides of t he fourth segmen t a t the base

,yel low. The

furrow bordering t he scu t el lum is crenulat ed ; t he post - scu t el lum is closely,longitudinal ly

striated. The area on t he met anot um is furrowed down the middle and is finely andclosely

,obliquely st riat ed. The propleuraebear curved , not very s t ou t st riae. The narrowed

basal part of the pet iole is shining and Obscurely striated ; the nodose (and larger) apicalpart is opaque. The st igma and alar nervures are fuscous ; t he apical abscissa of the radiushas a sharp

,straigh t , Oblique slope. The four an t erior femora are for t he greater part ,

the hinder are ent irely, black ; t he four fron t tibiae are yel low ,lined wit h black behind ;

t hepost erior yel low, black a t the base behind,t he t arsi are yellow.

5 . Trypoocylon erythroz ona tum,sp. nov.

Nigrum ,abdominis medio rufo ; t arsis anterioribus albis ; alis hyali nis ; nervis st igmateque

nigris. 2.

Long : 15—16 mm.

Hab. Hulu le, Male Atol l , Maldives.

Scape of an t ennae brow nish beneat h. Face,clypeus

,and t he lower part of t he eye

incision densely covered wi t h silvery pubescence. Mandibles for the great er part ferruginous.The upper part of the front bears a narrow longit udinal keel ; t he lower is raised

,the

raised part becoming gradual ly wider t owards the apex. Thorax thickly covered wit h ratherlongish white hair ; the furrow below the tubercles is smooth ; t here is a shallow cent ralfurrow on t hebasal region of t hemedian segment

,w hich becomes gradually wider and deeper

towards t he apex ; the apical furrow is wide : t he sides have an Oblique slope. Legs black ;the basal two j oin t s Of the four anterior tarsi and t he calcaria are white ; there is anarrow pale band near the base of t he hinder t ibiae. The wings are slightly infuscat edat the apex. The apex of the petiole and the second and third segments are rufous.

6. Trypox ylon melanurwm, sp. nov.

Nigrum,t arsis, tegulisquefusci s , alis hyalinis , stigmate fusco. J .

Long : 10mm.

Hab. Mamaduw ar i,Mahlos Atoll , Maldives ; M in ikoi , Laccadives.

Antennae black , fuscous t owards the apex ; t he last j oint is as long as the precedingthree unit ed ; on the underside i t i s dilated a t t he base, narrowed towards the apex ; t hemiddle j oints are sligh t ly di lated beneat h. Fron t and vert ex closely punctured ; on themis a large area ; i t s upper pa rt is rounded and encloses t he lower ocel lus ; its lower is

HYM ENOPTERA. 5 5

more narrowed towards the apex ; the part occupied by i t is raised and reaches near tothe ant ennae

,where it is prolonged as a short, st out keel ; from t he j unction of t he two

part s a curved keel runs t o t he eye incision. The lower part of the eye incision, t he

face and clypeus are covered wi t h si lvery pubescence. PrO and mesothorax smoot h andshining. The base of the median segment bears s t ou t , longi t udinal keels , the two centralof which are more widely separat ed ; t he rest of the segment is stoutly

,irregularly

,

transversely striated ; t he basal hal f has two longi t udinal keels which form a central area,

rounded at the apex ; there are t w o longitudinal keels on the apical slope,w hich unite

near t he apex of the segment , this second area being more sharply poin t ed at t he apext han t he basal one. On t he base of t he mesopleurae is a wide, deep ,

crenulated furrow,

bisected above by a smooth , curved one. Legs black , the tarsi testaceous, darker at thebase. Wings hyal ine , the nervures fuscous. Abdomen ent irely black.

Fam. S c oli idae.

7. Eli s thora ci ca Fab.

This fami ly is represen t ed only by the above, a common Eastern species. I t s J' has

not yet been described with any degree of cert ainty (cf. Saussure , Ca ta loga s Specierum

gener is Scolia ,p. 188

,and B ingham

,Fauna of B ri t. Indi a , Hymen,

p. 99) but , in thecol lection

,there are some males which can only be connected with thoraci ca. They are

densely covered wit h longish cinereous pubescence ; t he sides of t he clypeus broadly fromtop to bottom

,the apex of t he pronotum , t he yellow ex tending lateral ly to near the

tegulae, two broad marks on the base Of t he scutel lum and a mark on the centre of thepost - scutel lum

,yel low. The hair on the median segment is denser than on the rest Of

t he thorax ; the oblique furrow on t he fla t apical half of the mesopleurae is distinct. Thewings are hyalin e

,violaceous in tin t a t t he apex ; the radial cel lu les smoky. Legs black ;

the apex of al l the femora,the outer side Of the tibiae and of the base of t he front

tarsi yel low. Abdomen violaceous - black ; the basal three segments above have the apicesbroadly yel low ; the black dorsal par t s are di lated in t he middle and lateral ly, the basalproj ect ions being narrowed at the apex ; t he fourt h segment is more narrowly yellow atthe apex ; the black band there is also dilated , bu t only s lightly in the middle.

Hab. Common in M inikoi , Laccadives ; one of from Goidu,

Maldives.

Fam.S phegidae.

8. S celiphron madr aspa ta num Fab.

H ab. Goidu , Goifurfehendu Atol l , and Fainu , Mahlos Atoll,Maldives.

9. S celiphron vi ola ceum Fab.

H ab. Hu lu le,Male At ol l, and Mamaduw ari , Mahlos Atoll, Maldives. M inikoi, Laccadives.

Bot h these species of S celiphron are of universal distribu t ion in theOriental region.

5 6 P. CAM ERON.

10. Bembex ma ldivensi s , sp. nov.

Long. 17 mm.

H ab. Goidu , Goifurfehendu Atoll, Hedufur i , Mahlos Atol l.and Hulule, Male At ol l , Maldives.

B elongs to B ingham’s Sect ion A,and comes near or ientali s and borrer i. 2. Ant ennae

entir ely black. H ead densely covered with long, white hair, the clypeus also with si lverypubescence

, t he edge of the pronotum narrowly , t he labrum and the mandibles t o thetee t h

,pale yellow. Thorax densely covered with si lvery pubescence ; the prothorax , ex cept

behind, t wo narrow lines on the cent re Of t he mesono tum , t w o broader

,shorter

,almost

unit ed,transverse ones near t heapex , the apex and sides of t he scut el lum ,

the apex of t hepos t - scutel lum ; a broad curved line on t he cent re of t hemedian segment ; the sides broadlyfrom near t he middle , the base Of the mesopleurae, the mark dilat ed at the top so as t oenclose the tubercles

,a large vert ical mark in t he cen t re , which becomes gradual ly wider

from the top t o t he bo t t om ,a smal ler

,more irregu lar mark on the apex behind

,a crescent

shaped mark on t he met apleurae behind the spiracles and the apex broadly —t he markbelow ext ending backwards beyond t he midd le—su lphur - yel low ; legs yel low ; all the coxaeand trochanters ; al l the femora broadly in front and behind ; a line on t he fore t ibiaebehind and a shorter one in front , black ; t he spines on t he fore tarsi are long, stou t ,and , ex cept t he basal one , black. The basal j oint of the fore tarsi is not very broad,and becomes gradual ly narrowed , but not much , t owards t he apex ; t he second becomesgradual ly narrowed

, on the outer side from the base to the apex,which is transverse and

three times the w id t h Of t he base ; t he two following are smaller and are not so broadat the apex compared with the base. The basal abdominal segment is broadly black att he base and apex ; t he two bands are uni t ed in the centre by t w o black marks whichbecome gradual ly narrowed t owards t he apex ; t he apices of t hesecond , third , and fourt h arebroad ly black ; the black bands dilated backwards in t he middle , and have, at their base,two black marks

,which are broader than long ; t he fi fth segment is black , with t he sides

broadly yel low ; t he sixt h is ent irely black , with i t s sides st raight. The ventral segment sare black

,excep t t he first and second a t the sides of t he apex ; the yel low marks t here

are obliquely narrowed t owards the apex .

The J is similarly coloured ; the clypeus is pale yel low ,black a t t he base and wit h

a broad band on the apex ; t he sevent h j oin t of t he antennae is sligh t ly di lat ed near theapex ; t he eighth has a st out t oot h

,which becomes gradually raised from the base to the

apex ; at the apex is a smal l one ; there is a blunt basal and a smal ler, sharper apicalone on the nin t h ; t he tent h is broadly di lated behind , as is also, to a less extent , theelevent h. The middle femora are serrate beneat h ; t he j oints of the fore tarsi are not somuch dilated , especial ly towards t he apex ; t he spine on the second vent ral segment islarger than usual ; i t s downward lengt h is dist inc t ly greater t han i t s length at the baseand it is curved and slightly hook ed a t the apex. The pygidium is sparsely and distinct lypunc t ured on t he sides ; the epipygium becomes gradually raised in the middle t o shortlybeyond t he cent re of t he segment

,there curves downwards

,the top slightly proj ect ing.

The last j oint of the hinder t arsi is t hickly fringed at t he base wit h stiff longi sh spines,which being longer a t t he base and apex of t he fringed part give it an incised ap

pearance.

HYM ENOPTERA. 5 7

l l. Bembex handli rschi , sp. nov.

Long. 1 1 mm. J .

H ab. Hulule , Male Atol l, Maldives.

This is one Of t he smal lest Of t heOriental species. It belongs to the group of Orien ta lis.

Antennae entirely black ; t he seventh and eighth segments are dilated t owards theapex on the underside ; the nint h has a smal l sharp Spine ; the t enth and elevent h arehollowed ; the t enth is narrowed at the base. H ead black

,thickly covered with white

hair ; t he clypeus , except a t the base,and t he apex in the centre

,the outer orbits

,the

labrum, t hemandibles , except at the apex, two small central and a longer, pyriform mark

outside below the ocel li , pallid yellow . Thorax thickly covered wi t h whit e pubescence ; twonarrow lines on the cent re of the mesonotum and probably a transverse apical one

,a

narrow line on the pronotum ,t he base and apex Of the propleurae, a curved line on t he

upper half of the mesopleurae in t he centre ; this line becoming dilated gradually , butnot very much

,t owards the cen t re, a smal l mark on t he base of the met apleurae, a large

Oblique one, narrowed at the base, on t he apex and extending on to t he apical slope Of

the segment ; the apex Of the scu t ellum and postscu t el lum and a broad curved int erruptedline on t he apex of t he basal slope of the median segment , pale yel low. Legs colouredlike the body ; the coxae , trochanters, the base Of the femora narrow ly and t he upperpart broadly and the lower part s Of the anterior pair and t he binder knees , black ; themiddle femora are serrat e , but not strongly ; t he j oints of t he fore tarsi are slender andnot much di lated towards the apex ; the spines are long and pale. The basal black bandon the first abdom inal segment is broadly and roundly dilated in t hemiddle at the apex ;there are two black marks, broader t han long , on the base of the second , two less distinctones on the third ; t he apical bands are dilated backwards in the middle ; the band onthe fi fth is unit ed by a narrow line t o the base ; the 6t h and 7 t h segment s are enti relyblack. There is no spin e on the second ven t ral segment ; t he last is broadly depressedon the either side of t he midd le ; the ou ter edge is curved and narrowed towards t he apex.Wings clear hyaline ; t he cos t a and stigma testaceous ; the nervures darker. Tegulae black ,l ined with black on the inner side.

Fam. Vesp idae.

12. Poli stes hebraeus Fab.

Hab. Hedufuri , Mahlos Atoll,Maldives.

The Maldives form is ma caens i s Fab.

13. Poli s tes s tigma Fab.

H ab. M inikoi , Laccadives. The common Indian varie ty.

14. Rhynchium M aldi vense, sp. nov.

Long. 1 1—13 mm.Hab. H ulule , Male Atol l, Goidu ,

Goifurfehendu At ol l , Maldives. M ini koi , Laccadives.G. 8

5 8 P. CAM ERON.This species agrees so very closely with R. a rgenta tum Fab., sec. Saussure , S tet t.Ent. Z ei t.

XXIII. p. 18 7 (of. Cameron , Ann. and M ag. Na t. H i st Dec. 1900, p. that i t mightreadily be mist aken for it. I t may be known from i t by t hepronot um behind being sharplykeeled, by the propleurae being bordered before and behind by flat , shining keels ; thethird cubi t al cel lu le at the bot t om is broader, being as wide there as the space boundedby t hefirst recurren t and t he second t ransverse cubital nervures

,whereas in argenta tum it

is not half t he length.

J . An tennae covered wit h a pale down ; t he hook is brownish , stout , sligh t ly curved ;wi t h i t s pedicle i t is ful ly longer t han the j oint. The front , t he eye incision, and thelower half of t he ou t er orbits are t hickly covered with silvery pubescence ; t he vertex withgriseous hair. The front and vertex are close ly and s t rongly punctured ; the space betweenthe antennae bears a narrow

, but dist inct , keel. The clypeus at its greatest width is nearlyas long as its lengt h ; above it is transverse and bordered with a yellow band ; the upperhalf is dist inctly punctured ; the lower is alu t aceous and impunctate ; the apex is broadlyand roundly incised ; its sides have an oblique slope. Thorax st rongly and closely punctured ,except on the apex of the meso and on the basal hal f of the met apleurae. The prothorax at t he base al l round is bordered by a sharp keel ; the furrow above t he middleof the mesopleurae is distinct ; t he base below is smooth. The sides Of the median segmentarebroad ly rounded and bear neither teeth nor spines ; i ts apex has an almost perpendicularslope and is almost transverse ; above in the middle i t is very slight ly developed ; it issmoot h , except round t he edges ; the central furrow is deep and distinc t on the upper half.Legs densely pruinose

,as is also the abdomen. The basal segment Of t he abdomen is clearly

separat ed from t he second.

The 2 has t he clypeus punctured al l over and i t want s the white l ine on the top ;i t is more convex above and is more broadly and , not quite so deeply, incised a t the apex ;t hemedian segmen t is more fully developed

,and in t he centre it is finely , transversely st riat ed.

The wings in both sex es are deeply violaceous.

ANTHOPHILA.

Fam. Ap i dae.

1 5. H a li ct -as minikoiens i s , sp. nov.

Niger,densegriseo piloso ; tarsis longe fulvo pi losis ; ali s hyalinis ; stigmate testaceo, nervis

fusc is. 3 et J .

Long. 8 mm.

H ab. M inikoi , Laccadives.

Ant ennae bla ck ; t heflagel lum t ending to brownish near the apex and covered w i t h a palepi le. The ocellar region is smooth and shining ; t he rest of t he vert ex and front closely anddist inct ly punctured. The face is round ly convex in t hemiddle and is clearly separat ed fromthe clypeus ; and is there almost impunct at e. Clypeus rat her st rongly, but not very closelypunct ured

,except a t t heapex , which is t ransverse; t heex t reme apex is depressed , brownish in

tin t and slight ly projects a t t he ends. The hair is gr iseous , long and moderat ely t hick ont he front and face; sparser and shorter on the other part s. Mandibles brownish in t he

HYM ENOPTERA. 5 9

middle. Mesonotum shining, minu t e ly punct ured ; in the centre Of the basal hal f i s a shal low,

narrow furrow. Scut ellum punctured on t he sides and apex. Post - scutel lum t hickly coveredwith long griseous pubescence. The st riae on t he basal area Of the median segment areirregular ; at t he base they are stou t er and longitudinal ; at t he apex finer, closer and transverse ; ou the sides they are narrow ,

close together and longitudinal. The hair on the pleuraeis longer and thicker t han

_

i t is on the upper surface. The hair on t he legs is long andwhit e ; on the outer side Of t he t ibiae it is shorter and darker, almost black ; on t he tarsi itis long and bright fulvous ; the claws are pale tes t aceous. Wings clear hyaline

,iridescent ;

the stigma is testaceous ; the nervures are blackish ; the first recurrent nervure is in t ers t i t ial. Abdomen shining ,

impunctate ; the segments are edged with wh i t e pubescence ; thereis a narrow transverse fu rrow on the second and third segments ; the anal rima is brownish.

The eyes distinctly converge above ; the head is not very e longate in front ; the lowerhalf of the front bears a dist inct keel ; the apex of t he median segmen t has a semiperpendicular slope , is smooth and shining and deeply furrowed on the lower part ; thetegulae are piceous.

16. Cera t ina vir idi ssima (Dalla Torre) (vir i di s G uer.)

Many of the specimens are golden above.

Hab. Common in the Maldives.

17. Cera t ina appendicula ta , sp. nov.

Nigra,capite t horaceque supra viridis ; abdomine late flavo- lineato ; clypeo flavo,

nigrobilinea t o ; alis hyalini s , nervis st igma teque test aceis. 2.

Long. 4 mm.

Hab. Mamaduw ari,Mahlos Atoll , Maldives.

Scape of antennae broadly yel low beneath ; t he flagel lum brownish. Head dark green,

smooth,thickly covered wi t h white pubescence ; the clypeus and an Oblique, somewhat conical ,

spot close to its lower side,yel low ; in the cent re of t he clypeus are two large marks,

longer than broad and rounded above. Mandibles yel low , black at the apex. The mesonotum and scutel lum are green ; t here is a broad yel low line on t he pronotum

,extending

to the t ubercles,which are similarly coloured

,as are also the post - scu t el lum and a curved

line at the side of the scutel lum. The median segment is black and is, a t the base, closelyobliquely stri at ed. Wings clear hyaline ; the stigma and nervures are pale testaceous ; thesecond and firs t transverse cubital nervures are united at t he top. Legs thickly coveredwi t h long

,whi t e hair ; the knees, tibiae and tarsi are brigh t yellow ; t he middle t ibiae are

lined behind wi t h black in t he middle ; t he hinder are black , except at the base and apex.Abdomen smooth and shining ; there is a short ye llow line on the first segment on eit herside; on the second and third the yel low lines extend nearly to themiddle ; on the fourt hthey are almost united ; on the fift h and sixth t hey are continuous.

Characterist ic is the union Of the first and second cubi t al nervures a t t he top ; thet hird is straight and Oblique to near the bottom

,where it curves sharply backwards

,thus

forming a sharp angle,from the apex of which a short nervure issues ; the second recurrent

nervure is received a t the base of t his angle.A distinct species.

60 P. CAM ERON.

18. Allodapep i ci tar si s, sp. nov.

Niger, nit idus, facie , clypeo, maculaque inter antennas flavis ; alis hyalinis , stigmatefusco. 2.

Long. 5—6 mm.H ab. M inikoi

, Laccadives.

Antennae black ; the flagel lum wit h a brown ish tinge towards the apex. Head smoothand shining ; the vertex is Obscurely punctured. The front ocel lus is surrounded by a wide

,

deep furrow ; the front is broadly di lated in the middle ; t he dila t a ted part becomesgradual ly wider towards t he ape x . The yel low mark on the face is sligh t ly and roundlyincised on t he top ; the sides curve slightly outwardly ; the yel low mark is continued tot he end of t he clypeus, on which it is narrower and slightly longer. On the apex Of thepronotum is a broad, white band, narrowed a t t he sides ; the tubercles and the great erpart Of t he tegulae are yellow. Mesonotum and scutellum aciculated ; the bas al ar ea of themedian segment is st rongly and closely punct ured ; the rest of i t is smoot h and shining.Legs t hickly covered wit h long, whit e hair ; the hair on t he tarsi is more test aceous in colour

,

and their apices are test aceous. Abdomen shining,smooth ; t he apices of t he segment s

are testaceous ; t he basal segment at t he base is hol lowed,but not deeply. Wings clear

hyaline; the stigma fuscous ; t henervures slightly darker.

19. Megachi leotri ades, sp. nov.

Long. 13 mm. 2.

Hab. H i t adu,Mahlos Atol l

,Maldives.

Has the hair colorat ion of M . lana ta ,wit h which i t is closely relat ed ; it is smal ler ;

t he mandibles are different ly formed,their second tooth is longer compared t o the first

and not so widely separat ed from i t ; the middle part is broadly and distinct ly dilat ed ,t he posterior part being also di lat ed and separat ed ; i t is more rounded than t he middleportion , which has t hebase and apex s t raigh t and oblique.

Scape Of antennae black,opaque

,closely punct ured and covered with pale hair ; t he

flagellum is brownish. Head as wide as t he t horax , closely, almos t rugosely, punct ured exceptfor a T- shaped mark on t he clypeus and face, t he end Of which ex t ends to t he apex of t heclypeus. The fron t is t hickly covered wi t h long

,dark

,rufous hair ; the vert ex with short er,

dark,fuscous hair ; t he face and clypeus wit h long, dark hair. The mandibles are furrowed

along the edge t o near t he base ; on t he inner part t hey bear long,curved punct ures ;

the out er t oot h is sharply point ed , becoming gradually narrowed to t he apex , which isrounded ; the second t oot h is t riangular ; t he inner is broadly rounded ; the middle andthe second are bordered behind by a smooth , fla t part , which is clearly separated. Thoraxclosely and distinct ly punctured and t hickly covered wi t h dark, ru fous hair ; t he sides andbreast with shorter pale hair. Legs t hickly covered with pale hair ; t he hair on thebinder

tarsi on t he inner side is dark rufous. The basal segmen t s of t he abdomen are coveredw it h dark rufous

, t he ot her segment s are fringed w it h pale, fu lvous hair ; the scopa ispale , almost whi t e, except on t he last segment, where i t is black.


62 P. CAM ERON.

st raight on t he inner side, rounded on the ou t er, and narrowed beneath ; the lower t oo t his broadly rounded ; the lower somewhat triangular. The labrum in the middle is smoot hand shining

,bare and broadly keeled

,the keel being broadest on t he t op. The sides and

base of the mesonot um are closely punctured,as is also t he scu t el lum , except a t the base.

The median segment is more closely'

and st rongly punct ured ; t he basal area is clearlydefined ; its base is longer than the sides ; i t is aciculated and Obscurely furrowed downthe middle. The hair on t he legs is long , dense and deep black. Wings uniformly deepviolaceous

,with some blue tint s a t the base and apex ; the third transverse cubit al nervure

is obliquely sloped at the top and bo t t om ; the two parts are of equal lengt h and forma sharp angle in t he middle ; t he lower part has a more sharply Oblique slope. Abdomensmooth and shining ; sparsely punctured , except at t he base where the punctuat ion is closerand stronger ; t here is a smooth furrow down the middle of the apical segment ; themiddleOf the last apical segment is smoot h and keeled t owards t heapex.

This species comes near X. ga rdineri , but is not, I feel sure , i t s 2 ; as, apart fromthe difference in the coloration of the base of t he wings, it differs from it in somestructural points ; the area on t he median segment is longer , much broader compared toi t s length and it is glabrous and impunct ate ; t he thorax wan t s t he white pubescence ;it is smaller and t he abdomen at the base is not hol lowed : i t is distinct ly longer thant hehead and thorax uni t ed. The spine on t he outer side of the four hinder t ibiae is broadat t he base , bluntly pointed a t t he apex ; t hat on t he front pair is smaller

,more curved

and more sharply pointed.

23. Xylocopa aestuans Lin.

H ab. Common in the Maldives.

24. Xylocopa la t ipes Drury.

Hab. Mamaduw ari , Mahlos At ol l,Maldives.

A common Orien t al species.

25. Xylocop a gardiner i , sp. nov.

Nigra, nigro pilosa, alis violaceis, basi hyalinis. J .

Long. fere 25 mm.

Hab. Mamaduw ari,Mahlos Atol l

,Maldives.

Scape of antennae not dilated t owards the apex,slender. H ead Opaque , thick ly covered

with short black hair and closely and dist inctly punct ured,except a t t he sides of the binder

ocell i and the apex of the clypeus. Eyes not converging"

on the top,paral lel. The front

ocel lus is surrounded by a deep furrow. Clypeus fla t ; t he oblique lateral furrows are notclearly indicated. The base Of the mandibles is sparsely punct ured ; t he apical toot h isbroadly rounded a t t he apex ; it is short , not much longer t han broad. Labrum obliquelydepressed , closely and rather st rongly punct ured and wit h a smoot h furrow down the centre,which is broadest at t he top. The sides and base of the mesonotum are covered wit hpalehair ; t he cent re is bare. Except in t hemiddle t he mesonot um is closely and minu t elypunct ured. The scu t ellum is sparsely punctured a t t he base

,closely and more strongly

HYM ENOPTERA. 63

towards t he apex. Pleurae and sternum closely punctured and t hickly covered with blackhair. Wings deep violaceous, except at t he base , where t hey are narrowly hyaline. Legsthickly covered wi t h black hair ; t he outer side of t he hinder t ibiae is smooth and project sin t he middle int o a t riangular t ooth ; t he inner spur is broadly, triangularly dilat ed a t

the base. Abdomen shining ; above sparsely punctured at the base, much more closely a t

the apex. The last ventral segment is bare and smoot h and bears a stout t ooth on eitherside

,t hese teeth are st raigh t on t he inner side , rounded and narrowed towards the t op on

the inner ; below it is smooth and depressed a t the apex. The cubitus beyond the secondrecurrent nervure is straight , i t t hen turns up straight and obliquely towards the apexOf the cel lu le ; it t hen turns obliquely backwards , this upper abscissa being shorter and

having a more rounded curve ; the second recurrent nervure has the upper abscissa sligh t ly,t he lower (and longer) one is more sharply, Oblique. The basal area on the mediansegment is smal l , opaque , sparsely punct ured ; i ts length is as long as its wid t h a t thebase.

LAND CRUSTACEANS.

B Y L. A. BORRADAI LE , M.A., Lec turer in Na tur a l S ciences of Selwyn College,

Cambr idge.

(Wit h Plate III. and text - figures 12

CONTENTS.

Int roduct ion.

TheLand Crust aceans of M inikoi.

1. Genera l : t he spec ies and their hab i t a t s.2. Somepo int s in t he s t ruct ureand hab i t s of t he land hermit - crabs (Coenobi ta ).i. E x terna l fea t ures : st ruct ure and funct ion : systema t ic.i i. TheAl iment ary Cana l. A. The fore-

gut : deglut i t ion. B . Themid-

gut and l iver.C. Theh ind-

gu t and ana l va lve.i i i. TheVas c u la r System.

i v. The B rea th ing Organs : t he t hree resp ira t ory regions and t heir blood supply : the

mo isten ing of t hegi l ls : t hemovement s of t hescaphogna t h i te.v. TheK idneys.v i. TheNervous System.v i i. TheGenera t iveOrgans.v i i i. Reprodu ct ion : t heeggs and young.

Notes on t hehab i t s : l ifeunder wa ter, food, shells, sound- appara t us, et c.

3. Somenotes on the land crabs of t hegenu s Ocypode.i. Ocypodecera tophtha lma.

i i. Ocypodecordima na.

III. A l ist Of t he land and fresh - water Cru s t aceans collec ted in t heMa ldiveIslands.

LAND CRUSTACEANS. 6 5

I. INTRODUCTION.

I n t he economy of a coral island no group of animals is Of greater importance, fromthe biological point of View , t han the land Crustaceans. Their numbers

,their ubiquity

,

their act ivity combine to give t hem a prominence, w hich is all t he more marked fromthe absence of so many of the ot her

,land animals of continen t al areas. They are t he

chief scavengers Of the island , play a great part in the destruction or disintegrat ion of fruits,and probably aid in the dis t ribution Of seeds. The work done by them in burrowing alongthe sandy lagoon shore has a possible importance not hit hert o noticed ‘. And it is likelythat their omni vorous appetit e renders t hem enemies of many animals

,which canno t be

specified in t he present state of our knowledge : indeed it is highly improbable t hat theforegoing paragraph exhaus t s the list of instances in which their behaviour is a factor ofimportance in t he economy of the i slandz. NO st udy of a coral island

,in fact

,wou ld be

complete without an account of this cons t it uent of its fauna.

For five weeks,in the mont hs of June and July, 1899, I was with Mr Stan ley G ardiner

in t he Atol l of M inikoi. During this time I gave considerabl e at t ention to t he st ructureand habits of t he land crust aceans Of the island. My observat ions form the bu lk of thepresent paper

,but there is incl uded a repor t on the forms col lected by the expedit ion

in t he Maldives. In treating Of M ini koi I shal l begin wi t h a short enumeration Of t he

various Species t o be found there , and t he sit uat ions in which t hey respectively live,and

then pass on to some remarks on the struct ure and habit s of the land hermit - crabs

(Coenobi ta ), and the habits of the t rue crabs of the genus Ocypode. The report on t he

Maldive col lection ends the art icle.

In t he section dealing wit h the genus Coenobi ta I have thought wel l to enter intosome det ai l, and this for t w o reasons. First , that the peculiar habit s of t hese animalsgive them an int erest Of their own , and thus lend an import ance t o the investigat ion oftheir anat omy

,and secondly that , in Spite of several excel lent descriptions Of separate

syst ems of organs t o be found in the series of e legant and accurate works on the comparat ive anatomy of the Decapods which we owe to various French observers

,there is

sti ll need Of an account of the organization of a hermit - crab. In t hat Coenobi ta presentsthe Pagurine type in its most highly developed form (a t least as regards many of t he

organs), it is the most sui table genus for this purpose. In that it contains land animalsonly

,and is not found in Europe or temperate Nort h America, i t is less so. I have

endeavoured to overcome this difficu lt y by indicating, in t he course of t he article, those

points in which,to my knowledge , Coenobi ta differs from the hermit - crabs of t he sea

(Paguridae). My dissections have been made on fresh and spirit Specimens of the M inikoispecies 3 only, and principally on C. clypea tus and C. perla tus , bu t , in t he few cases whereot her information has been avai lable , i t has been used to check the dissections. Exceptin a few instances

,t he musculat ure has not been deal t with

,and histological det ai ls have

been entirely avoided.The in t roduction to this paper wou ld not be complete without an acknowledgment of

the generosi ty of the Drapers’ Company of London and Of the Managers of the Balfour

1 Seebelow,p.95.

2 Seee.g.Alcock, S ci.Mem.Med.Ofi.Ind., x n , p.5 9 3 Seebelow, p.68.G.

6 6 L. A. BORRADA ILE.

Memorial Fund , by which I was enabled t o visi t Ceylon and j oin Mr St anley Gardiner’sexpedition

,and t o undertake , among other pieces of work , that which is here set fort h.

My best t hank s are also due to Mr Stanley Gardiner for much kind advice and assist ance,and t o Mr Edwin Wilson for the care which he has best owed upon t he i l lust rat ions.

II. THE LAND CRUSTACEANS OF M INIKOI.

1. GENERAL.The dozen crustacean species living on land in t he Island of M inikoi more than make

up in number and act ivity for t heir comparat ive poverty in kinds. They are certainly t hemost conspicuous, as they are probably the most numerous, animals in t he island. TheCrabs are represented by six species only. Living in burrows, close along the high - watermark of t he sandy shore Of the lagoon , t he sage - green and yellow coloured Ocyp ode cer a

tophtha lma (Pallas) is very numerous. Inland , the place of t his species is taken “

by t he

brownish Ocyp ode cordimana Desm., whose underground galleries are plent iful in the sandysoil

,especially along pat hs and open spaces. Two species of Geograpsus—G.g rayi (H. M.

Edw .) and G.cr inipes (Dana)—~ are common , the former more so t han t he la t t er, which seemst o like damp spot s by the side of pools and t anks of fresh wat er for which G. grayi showsno great preference These act ive and conspicuous species, t hough they are rat her hardto dist inguish as sp i r i t specimens, are perfect ly dis t in ct in life , t heir colour alone servingt o separat e them

,were there no other differences. G. gr ayi gives the impression of being

black and whi te (in point of fac t t heback is purple and the legs and underside yel lowish),while G. cr inipes i s of a brigh t orange colour 1.

Anot her smal l G raps id belongs t o Kingsley’s subgenus 0r thograpsus 2, which is included

by Alcock 3 in t he genus Geograp sus, S t imps. Unfort unately, a part of the Minikoi collect ion was damaged on the way home

,and there are left of t his crab only two badly

mangled specimens. So far as can be seen from these,it is near t o, but not t he same

as , Dana’s Gr apsus longi tar si s“. The points in which the two forms differ are as follows

(fig. 12) —1. In the M ini koi specimens t he t eeth on t he underside Of

t he ou t er end of t he meropodi tes of t he walking legs are low and blunt,and much less marked than in Dana’s figure. 2. The same is the casewit h t he t oot h a t t he end of t he upper edge of t his j oint . 3. The hairyline along the dact ylopodi tes of t hewalking legs is want ing. 4. The toot ha t the ou ter angle of t he orbi t is shorter t han in G. longi tarsi s. It shie ldsonly about a fourt h Of t he cornea

,Instead of more than a half

,as in Dana’s

figure. The colour of t he crab when ali ve is a dul l brownish - green,and is

not much altered by preservat ion in spirit. The leng t h is abou t half aninch.

PIG - 12° sec ? ” In the absence of bett er mat erial , I am unwilling to give any opinionwa lk i ng leg of r i gh t

s ide of Geograpsmas t o the specific dist inct ness of t his lit tle crab. In the list below “

,i t

longi ta rs is’ var , mi wil l appear as a variet y (minikoiensi s) of Dana’s species.

m mu mThe last of t he crabs , t he l i t t le Metasesarma rousseaux i H. M.- Edw ., Is

1 For a careful discuss ion of t hedifferences between these 3 Alcock , Journ.As.Soc.Benga l, Lx rx .i i.3, p.394species , seeDeMan , Zool.J a hrb.Sys t.I x .pp.80if. 4 For references to l i tera t ureseebelow, p.68.

2 Kings ley, P.Acad.Phi lad.1880, pp.180, 194 if p.67.

LAND CRUSTACEANS. 67

found in various S i tuations,generally taking advantage of t he shel t er of some Obj ect. I t

Oft en hides under timber, where, as Dr Alcock remarks‘, i ts greenish mott led coloration is

protective. DeMan records 2 some examples of t his species from a stream in Flores , butit is cert ain ly not rest ricted to the water, nor even to particu larly damp Spo t s. With theexcept ion of Geograp sus longi tars i s , al l t he above crabs are included by Maj or Alcock amongt he Indian Fauna”.

The land hermit - crabs (genus Coenobi ta ) are no less numerous than the t rue crabs onM in ikoi. Three Species are found—C. perla tus H. M .- Edw ., C. rugosus H. M.- Edw . and C.

clypea tus Latr. Of these the first two Show a preference for the neighbourhood of saltwater

,while the third is chiefly t o be found in the j ungle. I shal l re t urn

,later on

,to

the subject of the genus Coenobi ta.

Three species of Isopod make up the tale of M inikoi land Crustaceans to t welve. Twowoodlice

,belonging respectively to t he genera Cubari s and Phi loscia , are found , as migh t

be expected,in rott en timber and loose ear t h , whi le a Ligia (L. ex oti ca Roux) lives in

certain localit ies along t he lagoon shore,and is chiefly conspicuous in running about on boats

drawn up on t he beach.

The fol lowing is a complet e systemat ic list of the land Crustacea of M inikoi

DECAPODA , B RACHYURA, CATOMETOPA.

Fami ly Oc ypodidae. G enus Ocyp ode, Febr., 17984.

Ocypodecera tophtha lma (Pa l las) , 1772.Cancer cer a tophtha lmus , Pallas , Spic i legia Zool. IX. p. 83 , pl. V. figs. 7 , 8Ocyp oda cera tophtha lma , Ortmann , Zool. J ahrb. Sys t. X. p. 365 (1897) [synonyms] ;Alcock

,Journ. As. Soc. B engal LXIX. ii. 3 , p. 345

Ocyp odecordimana Desm., 1825 .

Ocypoda cordimana , Desmares t , Consid. gen. Crust. p. 121 Ortmann , Zool.J ahrb. Syst. X. p. 362 (1897) [synonyms] ; Alcock, Journ. As. Soc. B engal LX IX.i i. 3 , p. 349

Family G rap s idae. G enus Geograp sus, S t imps., 18 5 8.

Geograp sus grayi (H. M. 18 5 3.

Gr ap sus gr ayi , H. Milne - Edwards,Ann. Sci. Nat. Zool. (3) XX. p. 170

Geograpsus gr ayi , Alcock , Journ. As. Soc. Bengal LX IX. i i. 3 , p. 395 (1900)

[synonyms].

Geograp sus cr inipes (Dana), 18 5 1.Gr ap sus cr inipes , Dana, Proc. Ac. Philad. 18 5 1 , p. 101.Geograp sus cr inipes , Alcock , Journ.As. Soc. B engal LX IX. i i. 3 , p. 396 [synonyms].

1 AlCOCk , Op.Ci t , p.428. that Fabr icius spelt th is nameOcypoda. In thecopies of the2 DeMan , Ma x Weber’s “ ReiseO.I.” II .p.850 E nt.Sys t.Supp l.in theCambr idgeUnivers i ty andZoologica l9 AJGOCk , op.Ci t “ pass im. Society’

s li braries thespel l ing i s Ocypode.Or tmann [Zool. J ahrb.Sys t. x . p. 359 s t ates

L. A. BORRADAILE .

Geograp sus longi tarsi s (Dana), 185 1 , var.mini koiensi s n. (fig.Gr ap sus longi tarsi s ,

Dana,Proc. Ac.Philad. 18 5 1 , p. 249 ; US . Ex pl. Ex pd. Crust. I.

p. 339, pl. X'X I. fig. 4

Or thograp sus longi tarsi s , Kingsley, Proc.Ac. Phi lad. 18 80, p. 195 .

G enus Metasesarma , H.M.- Edw 18 5 3.

Metasesarma rousseaux i H.M.- Edw 18 53.

Metasesarma rousseauaoi , H. Milne - Edwards,Ann. Sci. Na t ,

Zool. (3) XX. p. 188De Man

,Zool. J ahrb. IX. Syst. p. 138 Alcock

,Journ. As. Soc.

B engal,LXIX. i i. 3 , p. 427

ANOMALA ,PAGUR INEA.

Family Coenobi t i dae. G enus Coenobi ta , Lat r., 1826.

Coenobi ta perla tus H.M.- Edw ., 1837.

Cenobi ta perla ta ,H. Milne - Edwards

,H ist . Nat. Crust. I I. p. 242 Id. At las to

Crust. Cuvier’s R.An. pl. XLIV. fig. 1.Cenobi ta per la tus , Ortmann , Zool. J ahrb. Syst. VI. p. 3 19 (1892) [synonyms].

Coenobi ta r ugosus H. M .- Edw .,1837.

Cenobi ta r ugosa ,H. Milne - Edwards

,H ist. Nat. Crust. II. p. 241

Coenbi ta rugosus , Ortmann , Zool. J ahrb. Syst . VI. p. 3 17 , pl. X II. fig. 22 (1892)[synonyms].

Coenobi ta clypea tus Latr., 1826. (Fig.Coenobi ta clypea ta , Latrei ll e , Fam. Na t . R.An. p. 277Coenobi ta clypea tus , Ortmann , Zool. J ahrb. Syst . VI. p. 315 ,

pl. XII. fig. 20 (1892)[synonyms].

I SOPODA , ON IS CO IDEA.

Family A rmadi lli di idae. G enus Cubari s , B rand t , 1 833.

Cubar i s niur inus B randt , 1833.Cubar i s mur inus , B randt , Consp. Oni sc., B ul l. Soc.Na t . Mose. VI. p. 190 (1833)A rmadi llo mur inus , Buddes Lund

, Crus t . Isop. Terrest r. p. 27. Haun iae,

[synonyms].

G enus Phi losci a , Latr., 1803.

Phi losci a , sp.

Near P. gra ci li s B udde - Lund , 1879, but cannot be more accurat ely determined ,as t he single specimen now available for examination is in a somewhat damagedcondit ion

,and has lost bot h uropods.

Family L i gi i dae. G enus Ligi a ,Fabr., 1798.

Ligi a ex ot i ca Roux, 1828.Ligi a ex ot ica , Roux, Crust . Médit. 3 , pl. XIII. fig 3 Budde - Lund

, Crust. Isop.Terrest r. p. 266. Hauniae, 188 5 [synonyms].


70 L. A. BORRADAILE.

The fore-

par t of thecephalothora x is st rongly compressed , having a narrow ,vau l t ed roof

and fla t vert ical sides. I t s compression ex t ends also to the antennae and the first t hreepairs of legs, enabling t he whole complex , including t he l imbs

,to be withdrawn into shel t er

when necessary. The abs ence of t he rost rum is, no doubt , an adapt ation t o t he sameend. The cuticle over the whole of these exposed st ruc t ures is s t rongly calcified.

The eyes ta lks are narrow and elongated and directed s t raigh t forwards above t he antennules ; t he eyes themselves are small. Some rat her abnormalfeat ures appear in t he an tennules (fig. By the lengt hening oft heir j oints

,the stalks of t hese appendages far ou t reach t hose

of t he an tennae,bearing at their ends each a long and a short

flagel lum ,t hewhole being able to be folded up lengt hwise and con

cealed under t he body. The long flagellum,which is t he more

dorsal of t he t w o,is compressed and club - Shaped , wit h a blunt,

rounded end. I t s lower edge is covered,in t he dist al two - thirds ,

wit h a close fur of short “ olfact ory hairs‘. The short flagel lumt apers a t the end

,and bears on t he penult imat e j oint a long

F IG.14. Left An tennules. brist le, which looks as if i t were adapt ed for cleaning t he olfact ory

A E upagw us bernhardus ;hai rs of t he long flagel lum. The opening of t he ot ocyst is small

,

B. Coenobi ta clypea tus. bu t situat ed in i t s usual posi t ion on t he dorsal side of t he basalj oint of t his limb. The an t ennu les of t he Pagurids (fig. on

the other hand , are relatively short , and bear each two short flagella,like t hose of t he

crabs. The dorsal (outer) flagel lum tapers t o a point and bears below a fringe of ol factoryhairs much longer than those of Coenobi ta. These peculiari t ies of Coenobi ta are perhaps in connect ion wit h t he

habit of exercising t he sense of smel l in the search forfood 2 and tha t in a different medium from the one inwhich t he Paguridae live , while at the same t ime it isnecessary that the organs be able to be wit hdrawn intoa small space.

The scale of t he antenna (fig. which in t he

Pagurids is wel l developed and moveable , is, in Coenobi t a ,

reduced in S ize and ,excep t in one inst ance3, fused w it h

t he second joint of t heflat tened st alk of t he limb. Theflagel lum ends blun t ly ins tead of t apering to a pointlike t hose of most marine Decapods—a feat ure , whichmus t probably be at t ribut ed t o the same cause , whichhas Short ened the sensory hairs of t he an t ennule. The

b Tia};L

g“

cf'

e

E

z‘

faflur

é“ mouth- limbs (fig. 16) presen t no remarkable features.

dgnobi ta perla tus. 1. Antenna}; 26 5 81 They are all st out and Show a tendency t o develope

Tuberclewi th Open ing of green gland (not t uft s and fringes of st rong,close set, hai rs. The powerfu l

Been in A) cu t t ing edge of t hemandible is not t oo t hed , and t he lashof t he exopodit e of the first maxil liped is wan t ing.

1 Thestatement of Ortmann [B roun’s Thierrei ch, v. 2, mat ted in theabsenceof a support ingmedi umsuch as w ater ,

p. 1146] tha t theseappendages “ langen S inneshaareen t i s no doubt an adaptat ion to terrest ria l l ife.behren ” is thus somewha t mis leadi ng, t hough s t rict ly t rue. 2 Seebelow , p.92.Theshortness of theha ir s , wh ich would be apt to become 3 C.clypea tus, La t r.


The legs (figs. 13 and 17) of the first pair are unequal, t hough both are chelat e. The

larger, left chela, used for seizing and holding t he food and as a weapon of offence anddefence, serves, when the animal is withdrawn int o t he shell

,as a very perfect lid or oper

culum, for which purpose i t s st ou t , rounded Shape is c learly adapt ed. I t may also (0. r ugosus

A , fig. 17) bear a stridulat ing organ in t he form of a series of paral le l ridges on i t s ou t ersurface. The smaller chela is used for t earing t he food and conveying t he fragmen t s tothe mou t h. The second and t hird pairs of legs are adapted for walking.

The hinder p ar t of the thor a x , forming t hemiddle of t he three regions al luded to above,

may be said to begi n at the leve l of t he transverse port ion of the cervical groove and toinclude t he two hinder pairs of legs. I t is distinguished from the preceding re

gion by

being rat her depressed t han compressed and by less st rong calcificat ion of its in t egument .The branchiostegi tes , which anteriorly share in the general compression and calcification ofthebody

,are here soft and tend to be depressed. B ut t he legs of the fourth pair are carried

pressed up against t he sides of t he body in such a way t hat the soft branchios tegi tes areindented by t hem and only project dorsally and posteriorly where t hey overhang t he limbsin quest ion. B ehind , the branchiostegi tes gape widely from the body, and leave an opening ,through which t he legs of t he fi ft h pair can be thrust into the gil l chamber.

FIG. 16. Mou th - l imbs of the left s ide of Coeno

bi t a clypea tus from beh ind. A. Mandible; B . l s t

max i l la ; C. 2md max i lla ; D. 1s t max i ll iped ; E. 2ndmax illiped ; F. 3rd max il liped. A sma l l port ion of

theanterior side of D i s at t ached to show theendo

podi te.In Pagurus themax i l l ipeds bear on their ex opoditeslong, s lender , curved la shes, fr inged wi th fineha irs andmuch resembl ing thoseofmany crabs. Theabsenceoftheses truc t ures in Coenobi ta.i s perhaps duet o thelossof their funct ions in theabsenceof wa ter. Theli t t lebackward hook on the firs t max i l la is wan t ing i -n

Pagurus.

72 L. A. BORRA DA ILE.

The four th and fifth legs (B —E,fig. 17 ) of each side are markedly smaller t han the

preceding ones. Each of them is chelat eaft er i t s own fashion

,and each bears a

part in the work of holding t he animalint o i t s shel l. This is accomplished bymeans of a definit e pat ch on the propodite of t he limb , which is covered witha fine sculpture in the form of roundedscales, overlapping one another so as toproduce a surface capable of friction. Incases, which often occur, when t he animal

,

eit her from its size , or from the lack ofsuit able gastropod shells

,is forced to con

ten t i t se l f w ith some other receptacle forits abdomen ‘, t hese limbs have t o t akea larger share of t he work in which theyordinarily assist the sixth abdominal appendages , of keeping on the “ house. Itis probably by their means also that t heanimal is able to emerge from the shel l

F IG. 17. Legs of hermit - crabs from left s ide. A. ls t W hen i t is invert ed. The Chela' Of t heleg (chel iped), Coenobi t a rugosus ; B. 4th leg, E up agurus bern four th leg is of a, peculiar nat ure , Theha rdus ; C. 4th leg, Pagur

'us deformi s ; D. 4th leg, Cjoenobt ta pl

‘

OpOdlte, on which the above - mentionedclypea tus ; E. 5 th leg, Coenobt ta clypea tus. 1. S t r idulatmgorgan. 2.Malegenerat iveopening. pat ch 15 developed

,IS a broad

, dlSCOlClal

st ruct ure. Against this disk fits a lit tlesickle - shaped dactyle. In Pagurus the limb more nearly approaches a normal chela

,in

Eup agurus it is sub - chelat e. The fifth leg has a chela which is clumsy, but of the ordinaryshape

,with t w o approximately equal fingers. As in the other forms (Anomala) in which i t

is a gil l - cleaning organ , it is covered wit h hairs and usually carried under the branchiost egi t eduring life

,though upon occasion it can be used t o assist in retaining hold of t he shel l.

The abdomen is connected by a narrow waist w i t h the t horax. I t is covered wi t h asoft

,fle x ible skin , save for a narrow t ransverse ridge represent ing the t ergite of each of

the first five abdominal segments,and broader plates on the sixth segmen t and telson.

The whole is spirally twi st ed t o t he righ t , t o fit in wi t h t hedex t ral twist of an ordinarygastropod shell. In correspondence w it h t his twist , the appendages are also asymmet rical.In the male t hey are absent from the first five segments 2, but in the female t he second

,

t hird and fourth are provided , on t he left side wit h long, biramous limbs, covered wi t h hairsand used in t he breeding season for carrying t he eggs. The sixth pair of abdominal limbs

(fig. 18 ) are present in bot h sexes , but t hat of the right side is larger than t hat of the left.The exopodit e is S ickle - shaped and longer t han t he endopodite ; both are prov ided , on theou t er surface, w it h friction organs like those of t he last two t horacic legs, and the funct ionof t hewhole limb is obviously to anchor the animal into its shel l.

1 e.g. theshel l of the fruit of Ca lop hyllum or of the 2 C.clypea tus has short , uniramous limbs on theleft s idecoconu t. Seep.92. in the2md, 3rd and 4th abdominal segment s of themale.


The family Coenobi t idae, containing besides Coenobi ta the Robber - Crab B irgus , differsfrom the Pagurids, or hermit - crabs of the sea ’, in the fol lowing ex ternal feat ures —1. The

F IG. 18. Last t wo abdomina l segment s of Coenobi ta clypea tus. A. Dorsal view ;B . Ventra l view. 1. Valveguarding theanus.

st ructure of the antennules (seeabove). 2. The st ructure of the ant ennae (seeabove). 3. Thegreater compression of the fore—part of t he body. 4. The habitat—ou land.

The genus Coenobi ta.differs from Bi rgus in the form of the abdomen,which in B trgus

is short,untwisted

,provided w i t h broad terga, and not carried in a shel l. The poin t s in

w hich the outward features of Coenobi ta vary from species to species are smal l and nu

important. The most int eresting is perhaps t hat which separates C’. clypea tus La tr. fromt he rest of the genus. In this species t he vestige of the ant ennal scale is s t i l l looselyart iculat ed with the st alk. In al l t he o t hers it is fused. C. di ogenes (Catesby) is dist in

gui shed by its cylindrical eyes talks from the other species , in which these st ructures arealways compressed. Of the remaining four forms , 0. rugosus H.M.- Edw . and C. per la tus

H. M.- Edw . are dist inguished from C. comp ressas Guérin and C. spinosus H : M.- Edw . by t heelongation of the basal j oint of t he fi fth leg of t he male into a geni t al process

,and by

the presence of stridulat ing ridges on t he outside of t he hand (propodite). The smal lgrey or purple C. rugosus is very variable (Bouvier

2 enumerates several varie t ies), but maybe easily dis t inguished from the larger, scarlet, C. per la tns by the grea t er development inthe latter of t he genital process. 0. compressa s is recogni sable by i t s smal l size , grey, orpurple colour and less hairy integument from C. sp inosus, which is larger, of a brown colour,and considerably hairy in parts. The last revision—and the one adopt ed in thi s paperis to be found in Or tmann’

s wel l - known work on t heDecapods of the St rassburg Museum 3.

It differs in several points from that of Bouvier“,but has t he meri t s of simplicit y and

clearness, and has satisfactori ly accounted , so far, for every one out of t he many specimensthat have passed through my hands.

The distribu t ion of t he genus would appear,from t he data in Ortmann

’

s paper andothers since published

,t o be as fol lows—C’.diogenes is found in the West Indian region

alone,and is t he only represent ative of the genus t here. The other species are all re

s t ric ted t o the Indo - Pac ific area, but are there of almost universal distribu t ion , 0. rugosusbeing perhaps t hemost widespread.

1 Seefootnoteto p.69 above. 3 Or tmann, Zool.J ahrb.Sys t., v1.p.3152 Bouvier, B ull.Soc.Phi lem., Par is , II.p.143 Bouvier, loc.c i t.G.

7 4 L. A. B ORRADAILE.

i i. The A limen t ary C ana l.

A . T he Fore- gu t‘. (Pl. III. figs. A—E). The short, wide oesophagus presents no

feature of special int erest. In correspondence wit h the compression of the fore - part of thecephalothorax mentioned above fi, t he Maw or S tomach is narrow and elongated. Thehinder or pyloric division is nearly horizon t al

,and is sundered from t he an t erior “ cardiac ”

portion by t he usual deep hol low on t he dorsal surface of the st omach. The ca rdi ac sac

is narrow,rounded in front and somewhat flat tened above. The wal ls of i t s t hin fore - part

Show on each S ide t he t hickened “ cardiac disk ” found in ot her Pagurinea. (8 , Pl.

Int ernally , this disk bears seven or eight t uft s of long hairs, placed on pointed prominencesarranged along i t s lower edge.

The gastr ic mi ll is strong, no doubt in connec t ion with t he nat ure of t he food , whichis chiefly veget able

,and oft en consist s of such t ough subst ances as the fruit s of t he screw

pine (Pandanus). In t hemesoca rdi aeossi cle(1 , Pl. the mos t prominent part is a strongband across the roof of t he cardiac division of t he s t omach , ben t in t he form of a bow ,and lying with t he hol lowed side forward ; i t s ends are broadened. In fron t of t his is at hinner port ion , which merges gradually int o t he t hin wall of t he fore - part of t he st omach.B ehind t he ossicle t he roof of the st omach falls away int o the hol low between t he twodivisions. The rounded hinder edge of t he bow project s somewhat over t his gap , andoverhangs a wide , triangular process from i t s own underside, which forms part of t he hindwall of the cardiac port ion—t he ant erior wal l of the hol low. The apex of t his t riangu larplat e, by which it j oins t he fore end of the urocardiac ossicle , is not poin ted , but endsin t w o rounded lobes, of a bright whit e colour and st rongly calcified. The p terocardi ac

oss i cles (2, Pl. III.) are t ri angular or rat her t hree- lobed struct ures in t he side wal l of t he

cardiac division , at t he ou t er ends of t he how of the mesocardiac. One lobe is directedou t w ards, one inwards , and one downwards. B etween t he ou t ward and t he inward lobeslies that edge of the ossicle which art icu lates w i t h the broadened outer end of the mesocardiac how . The z ygocardi a c ossi cles (3 ,

Pl. III ? ) also lie in t he side walls of t he s t omach ,but behind t hemesocardiac. Each is a roughly diamond - shaped plate of large size

,placed

so that i t s longer diagonal runs lore and aft . There are t hus an an terior and a hinderangle. The lat ter is directed somewhat downwards

,so t hat

,of t he t w o edges of t he

diamond which are uppermost, t he hinder slopes downwards , while the more anterior is

almost horizont al. This la t ter edge is much t hickened and curled inwards t owards t he

cavi ty of t he st omach. The ot her edges a re also t hickened , t hough not to t he same ex t ent.The la ter a l tooth (35 , Pl. borne internally by t he zygocardiac ossicle of each side , iss t rong, and bears a series of t ransverse ridges which grow smaller from before backwards.The first t hree or four of t hese are much larger t han t he hinder ones , and are set

fart her apa rt .

1 In wr i t ing the fol low ing accoun t of the fore-

g ut and thet hreeM inikoi species. Thus thet ot a l mater ia l ava ilablei t s armat ure in Coenobi t a , I have been much helped byMocquard

’s work on theseorgans of t heDecapods [AH IL S c'i .

Na t. (6) Zool.x v1.i. andespecia l ly by thenumerousa l lus ions to th is genus sca t tered through h i s sect ion on the

Pagur inea. Though hefigures no Coenobi ta , Mocq uard had

dissected threespecies—C.comp ressas , C.sp inosus, and a th irdwhosenamehedid not know. Thepresen t account i s basedon thein forma t ion given by Mocq uard and on dissec t ions of

has reference to a t leas t fiveout of thesi x species of t he

genus.2 p.70.3 Th is oss iclei s so placed tha t i t overhangs and presen t s

onefacedownwards and an edge(thehori zon ta l one) outwards. I t s shapei s thereforenot properly seen in a full s ideview and cannot bemadeout in fig.A,

Pl.III.but may beseen in theven t ra l view (fig.C).


7 6 L. A. BORRADAI LE.

sharply marked off. The medi an tooth (28 , Pl. is large and well - developed. I t is anoblong plat e

,somewhat short er t han t he urocardiac ossicle, a t the hinder end of which i t

is placed,and wi t h concave sides. The inner surface bears a median longi tudinal and four

or five pairs of lat eral t ransverse ridges , of which t he t wo hindermost are t he s t ou t es t.The det ails vary considerably in different Species '. The p repylor i c ossicle(5 , Pl. III.) is aT- shaped s t ruct ure

,placed vert ically in t he fron t wall of the pylori c division of t he s t omach

,

i .e. in t hat wal l which faces the hol low between the two divisions. The s t em of t he Thas concave sides and broadens considerably , w here it j oins t he cross - piece. The lat ter issligh t ly concave on i t s ant erior side.

The oss icles desc r ibed in t he forego ing pa ragraphs const i t u te t he gast ric mill, and,wi t h t he

mu sc les wh ich move t hem,a re t he appa ra t us by wh ich t he food is t ri tura ted. To complete t he

account of t he skelet on of the st omach i t i s now necessary t o ment ion cert a in less import antp ieces, by wh ich t he organ ma int a ins i t s shape

,and wh ich support inter na lly t he va rious t uft s

and fringes of ha irs wi t h wh ich i t i s provided. W emay cons ider t hese in four groups

(1) Those of t he cardiac region, a ll of wh ich lie in t he vent ra l,or vent rola tera l wa l l.

They comprise t he cardiac disks (descr ibed above), t he post - oesophagea l brushes, t he cardiac s idepla tes

,t he prepec t ina te oss icles, t he combs

,subdent ary, penna te and inferola tera l ca rdiac oss icles

,

and theca rdiopy lor ic va lve.(2) Those on t he vent ra l wa l l of t he pyloric region ; compris ing t he auri cula r

,antero inferior

pyloric , preampullar, and posterior subampullar oss ic les.(3) Those a t the s ides of t he py loric reg ion ; compris ing t he anter ior, middle, and posterior

pleuropylori c and anterior and middle subampulla r oss ic les.(4) Those in t he dorsa l w a l l of t he pylor ic region ; compris ing t hemesopylor ic and uropy

lor ic oss i c les.G roup 1 . PostoesoPhagea l bru shes (29, Pl. A pa t ch of ha irs on ea ch s ide of t he stomach

,

s i t ua te on a lobe formed by invagina t ion of t he wa ll on t he vent ra l s ide, a li t t le beh indt he oesophagus. The ou ter open ing of t he invagina t ion i s surrounded by a ca lc ified ring.

Cardi a c si de-

p la tes (10, Pl. Large t h icken ings of t he vent rola tera l wa l l beh ind t he oesophagus.They a recovered interna lly wi t h t uft s of ha irs, espec ia l ly long t owards t he h inder end.

Prepec t ina te ossi cles (11, Pl. E longa ted pla tes in t he s ide wa l l of t he s t omach, above t he

ca rdia c s idepla tes. The fore end of each is t h in and ex panded and t he upper edge t h ickened a long i t s whole length , but espec ia lly a t t hebinder end.

Combs (30, Pl. An invagina t ion of t he wa l l of t he st oma ch a t t he h ind end of t he pre

pec t ina te oss ic le of each s ide, produc ing interna lly a cush ion on wh ich a rea number of short,

s t ou t spines.Pect ina teand post

-

p ect ina teossi cles a rewant ing.S ubden tary oss icles (14, Pl. A s lender bar runni ng backwa rd and downwa rd on each s ide

from t he zygoca rdiac oss icle t o t he foreend of t he anteri or subampulla ry oss ic le, wi t h wh ichand not wi t h t he inferola tera l cardiac oss icle, i t art icula tes.

Penna te and i nferola ter al ca rdi ac ossi cles (12 and 13, Pl. Tw o slender bars runn ing para l lelwi th one anot her a long t he l ine of junct ion of t he vent ra l and la tera l wa lls of t he st omach,wh ich line is defined by t heir presence. A t t hei r h inder ends t hey curve upwa rds

, in corre

1 SeeMocquard, 0p.ci t., p.131.


spondence wi t h the diminish ing height of the cardiac div is ion. Each of t hem grows gradua llybroader from before ba ckwa rds, and i s ex panded a t i t s posterior end. The vent ra l s urfaceof t he s t omach, bet ween t he tw o inferola tera l cardiac ossicles, i s t h ickened on each s ide int oa pla te of ca rt i laginous cons istency. In t he middle l ine t hese t w o pla tes a re sundered by a

st rip of t h in ch i t in,in wh ich l ies a median oss icle hav ing t he shape of a dagger, placed

lengt hwise wi t h t he blade forwa rds. The penna te oss icle, wh ich i s t he more dorsa l of t he

t w o,bears interna l ly a fringeof long ha irs.

Cardiopylor i c va lve (31, Pl. Th is has t he form pecul ia r t o t he Pagurinea and G ala theinea

t ha t i s t o say, i t bears a V- shaped eleva t ion composed of close- set lamellae not unl ike the

ridges of a file. The lamel lae a re flex ible and t he po int of t heV is backwa rds. '

The wholest ruct ure i s known as t he “ median inferior t ooth.

La tera l cardiopylor i c ossi cles are want ing.G roup 2. Aurieu la/r ossi cles (15 , Pl. A pa ir of s t out t roughs, runn ing backwards and

inwards from t he h inder ends of t he inferola tera l ca rdiac ossic les, on the vent ra l s ide of t he

st omach. The hollow of t he t rough i s t owa rds t he ins ide of t he s t oma ch, and the h ind endof ea ch i s ex panded and less st ou t t han i t s forepart . Thereareno a ur ic les.

An teroinfer i or pylor i c ossi cles. A semi - t ranspa rent,four s ided, median pla te wit h a t riangu la r pro

jec t ion from t he middle of i t s fore edge. Beh ind i t , and sepa ra ted from i t by t he innerends of t he preampullar oss icles , a re t w o sma l l, semi - t ransparent , t riangular pla tes, one oneach s ide of t hemiddle l ine.

Preampullar ossi cles. Semi - t ranspa rent , t ransverse pla tes, just in front of t he ampu l lae on t he

floor of t hepyloric div is ion.Posteri or subampu llar ossi cles (17 , Pl. S t ou t t ransverse bars, one on ea ch s ide jus t beh ind

t he ampu l lae. Their inner ends a reex panded and meet , but do not fuse.G roup 3 . An ter i or pleur opylor i c ossi cles (18 , Pl. Each of t hese s t a rt s a t i t s h inder end

by a rough ly t riangula r ex pans ion, wit h t he apex directed downwa rd and forward. Fromt h is apex proceeds a s t ou t r idge on t he s ide wa l l of t he s t oma ch

,wh ich

,runn ing forward

,

ends by a rt icula t ing wi th t he vent ra l end of t he subden t ary. The lower s ide of t he ridgebears a t riangu la r ex pans ion. In t he space bet ween t h is oss icle and t he pylor ic, t he wa ll of

t he s t omach i s t h ickened t o form a ca rt ilaginous pla te.M iddle p lenrOpg/lori c ossi cles (19, Pl. A s lender hori zont a l rod

,beh ind and below t he pre

ceding oss ic leon each s ide.Poster ior p leuropylor i c ossi cles (20, Pl. A. short , cu rved bar

,wi th t he convex ity forwards

,

beh ind t hemiddle pleuropylori c of ea ch s ide.An ter i or subampu llar oss i cles (21, Pl. An elongated ca lcified st r ip, hollowed on the outer

surface,ly ing above t he a u r icula r oss i c le of each s ide.

M i ddle subampu lla r oss icles (22, Pl. Th icken ings of irregular form,bea ring t o t he ampullae

t he same rela t ion t ha t t he anterior subampullars do t o t hea ur ic ular ossi c les.G roup 4 .

"

Uropylori c ossi cle (7 , PI. A broad median pla te in t he pyloric roof, wi t h t he

anter ior edgevery concave, and t heanter ior angles produced and st rongly ca lc ified.Mesopylori c ossi cles. The anter ior pa ir are want ing t heh inder ones (23, Pl. III.) arepresent a s

a sma l l ova l pla te on each s ide,in front of t he uropyloric and out s ide i t s produ ced fore

angles.

7 8 L. A. BORRADAILE.

I nterna lly t he pyloric div is ion present s t he usua l fi l tering appa ra tus of va lves and fringes wi thout any rema rkable fea t ures. Of t he termi na l va lves t he dorsa l and la tera l are

'

much elonga ted and po inted.It wil l perhaps be of interest here to indicat e t he bearing of the gastric armat ure

just described on the sys t ematic posi t ion of t he genus‘. The fol lowing charact ers belongespecially to t he fore - gut

,of the Pagurinea :—(i) The long, narrow shape of the st omach.

(ii ) The slight inclination of the pyloric region. (ii i) The form of the mesocardiac ossicle.(iv) The longitudinal direction of the dorsal edge of the zygocardiac ossicle. (v) The elon

ga t ion of the cardiac side plat es. (vi) The almost horizontal direct ion of the pennate andinferolat eral cardiac ossicles.

Within the Pagur inea , t he following characters separate the Coenobi t idae2 from t he

Paguridae : (i) The lower edge of the cardiac disk has i t s hairs arranged in t ufts onpoint ed proj ect ions of the body of t he disk. (i i) The lateral t oot h has no not ch on itslower edge. (i ii) The median tooth carries t ransverse ridges. (iv) The ex opyloric ossicleshave a s t rong sadd le - shaped curvature. (v) The pyloric ossicle is expanded and strengt henedat its an t erior angl es.

The differences which separate Coenobi ta from B i rgus are smal l , but the fol lowing maybe ment ioned —(i ) The anterior t ubercle of t he lat eral tooth is wan t ing in Coenobi ta. (i i) Thehinder edge of t he mesocardiac ossicle is rat her more convex than in Bi rgus. (iii) Theanteri or edge of t he cross - piece of the prepyloric ossicle is more concave. (iv) The calcifiedring round t he opening of the in vaginat ion on which the ‘post - oesophageal brush is borne,articulates, in B irgus, with an elongat ed t riangular strip in front of the pennate ossicle.(v) In B i rgus, the cardiac disk and the arrangement of hairs on i t s lower edge are moredeveloped than in Coenobi ta ,

and form t he “ suboesophageal valves ” of Mocquard3

The gas tri c muscula ture shows no import ant difference from that of t he Pagurids asfigured and descri bed by Mocquard. The st rong gast ric mi ll is provided with a correspondingly powerfu l set of muscles to work i t . The anter ior gas tri c muscles are a pair of st outstrands insert ed on t he mesocardiac ossicle and diverging slightly as t hey run forwardthence to t hei r origin from t he under side of a low,

rounded,transverse ridge of t he carapace,

sit uated above t he ba s es of t he antennae and eye - s t alks. This ridge is t he “ procephalicapophysis.” The posterior gastric muscle of each side is divided into t w o bundles—an

inner one , inserted on the t hickened plate at the outer end of the pyloric ossicle, and anouter one , somewhat broader t han the inner, insert ed on the ex opyloric ossicle. The originof t he inner bund le is par t ly from the ant erior side of a t hin , flat

,triangular apophysis ,

which proj ec t s inwards and somewhat forw ards from t he cervica l groove a short distancefrom t he middle line

,and part ly from t he carapace in front of this apophysis‘. The out er

bundle arises in fron t , and a lit t le to the ou t side , of t he other. The cardiopylor ic orsuperi or ca rdi ac muscles consist on each side of t hree bands running backwards from thehind edge of t he mesocardiac t o be at t ached t o t he ex opyloric ossicle. Of these three,

1 For thefact s on wh ich thefollowing paragraph i s based 4 Such an apophys is seems , from Mocquard’s remarks on

I am indebted to Mocquard (op.c i t ). thegas tr ic muscula t ureof Pagurus , to bewant ing in that2 Thefami ly Coenobi t idaeincludes thegenera Coenobi ta genus. I t i s however present in E up agur

-

as bernhardus ,

and B i rgus. wherei t i s directedforwards.3 Ann.S ci.Na t., XI I I.3

LAND CRUSTACEANs. 79

that which is outermost at its origin from themesoca rdiac ossicle passes under t hemiddleone and is at tached posteriorly bet w een this latter and t he inner band ’.

The cardiac disks are provided,as in t he o t her Pagurinea, w i t h di la tor muscles , running

forward to t he cephalic wal l. Mocq uardfi,discussing the working of the gastric muscu lature

of the Decapoda,concludes t hat t he funct ion of t he di lat ors of t he st omach 3 is , by en

larging its cavity, t o draw int o i t fluids,carrying wi t h them t he solid part icles of t he

food. On the relaxat ion of the dilat ors , the const rictors“ of t he stomach wil l drive outthe liquid

,while the solid matt ers wil l be caugh t on t he fi lter provided by t he hairs on

t he cardiac disks or elsewhere near t he opening of the oesophagus. This t heory is veryplausible

,but there are considerable difli cult ies in t he way of its accept ance in the case

of t he genus Coenobi ta at least . For the great maj ori t y of the food is by these creatureseaten absolutely dry

,with on ly such juices as i t natural ly cont ains. If, for inst ance, an

individual be wat ched in t he act of consuming one of its most common art icles of food—t he fru it of the Pandanus—i t wil l be seen to hold t he food fi rm ly wit h t he great chela

,

while the smaller one is employed in stri pping off t he fibres of t he fruit and placingthem between the max i l lipeds of the third pair, which open to receive t hem,

and t henpass them on towards the mouth. They are not immediate ly rej ected

,and presumably

are sent into t he stomach aft er having undergone a first crushing by the mandibles. Inany case t he powerfu l gastric t eet h argue a masticat ion of some part of t he food there.Now the Pandanus - fibres

,and indeed the maj orit y of the

_food -

of al l sort s, cannot besupposed to contain enough moisture to convey t he solid part t o t he s t omach

,however

much it may be broken up by the mandibles. And t he fact that the st omach,when

opened,contains but l it t le fluid

,prec ludes the suggestion t hat its watery cont ents , passing

backward and forward , con t inually perform t he same function of carrying sol id food from t he

mouth to the st omach.

These considerations appear to conclusively negat ive Mocquard’

s t heory of swal lowingin t he Decapods. For

,although the same difficu lt y does not exist in marine forms

,t he

similari ty of the mechanism in the t w o cases makes i t difficu lt to suppose that a differen tmethod is adopted in each. The subject is a t al l even t s wort hy of furt her investigat ion.It may prove t o be t he case that the more liquid part of t he food is swal lowed in theway described by Mocquard (and suggested before him by Parker 5 ) while the more solidportion is either rej ected or swal lowed by some ot her mechanism

,as

,for instance

,by t he

constrict ors of the oesophagus.

B . T he M i d -

gu t“. (Pl. III. figs. F

,G .) As in ot her Decapods , t hes t omach of Coeno~

1 Themodeof act ion of thegas t r i c mil l of Decapods i sdiscussed in Mocquard

’s paper and i n H ux ley’

s“ Inver te

bra tes.” B r iefly put , i t i s a s fol lows —Thecont ract ion of

thean ter ior and poster ior gas tri c muscles ha s theresul t ofbr inging thethreeteeth (median and la teral) together. On

their rela x a t ion theoss icles ret urn t o their norma l pos it ionby theelas t ici ty of thes tomach -wa l ls , and a lso part ly by theact ion of thecard-iopyloric muscles.

2 Mocquard, Ann.S ci . Na t., XI II.3,p.3 and x vr.1 ,p.25 5.

3 Of coursetheabove-ment ioned dila tors a t ta ched to thecepha l ic disks , though they arethech ief, arenot theon lyones. Set t ing as idethedi lators of theoesophagus and thoseof thepy lori c divis ion , thefol lowing, at tached on thecardiac

hal f of thes tomach,aredescr ibed by Mocq uard ; theantero

super ior (on thecepha l ic disk), the anteroinfer ior , the an

terior and thepos ter ior latera l.4 Of theset herei s oneon each s ide—a band of fibres in

serted on thecardiac wa l l at thefront endof thes ide- plate, andrunn ing downwards and a l i t t leforwards t o theoesophagus.

5 T.J .Parker , Journ.Ana t.Phys.n .p.596 B ouvier , in a shor t ar t icle Sur la Respira t ion et quelques disposi t ions organ iques des Paguriens terrest res du

GenreCénobi te (B ull.S oc’

.Phi loma th., Par is , (8) II.p.194remarks briefly on th is and other in terna l organs of

Coenobi ta. H is observa t ions weremadeon C.diogenes , andi t i s interest ing to find that , so far as they go, they supportmine

,wh ich relatet o theIndo - Pac ific species.

80 L. A. BORRADA ILE.

bi ta is fol lowed by t hemid—gut or mesenteron—a short, soft - walled tube, which grows gradual lynarrower from before backwards in consequence of t he difference in width be tween the

pyloric division of the st omach and the hind -

gut . Through the roof of t he mid - gut t hereopens on each side a short

,curved coecum of simple form wit h its blind end direct ed

forwards. Through the floor open the t wo bile ducts.

TheLiver (Pl. III. fig. F). The same compression of t he t horax , which has affect ed theshape of the stomach

,has brought about the displacement of the liver of the hermit- crabs

backwards int o the abdomen. Start ing from the underside of the mid - gut,one close on

each side of the middle l ine , t he bi le ducts may be traced backwards below the hindgut

,gradual ly diminish ing in calibre all the way. On i t s ou t er side each duct bears

,at

almost regular int ervals , about a dozen stout tubes or primary duct ules , the first of whichen t ers nearly at the level of t he las t thoracic leg. Each of these ductules curves upwardsand inwards over the hind -

gut , diminishing in diameter by giving off secondary duct ulesalong its outer side. These in t urn give off duct ules of a t hird order, and t he lat t er beart he terminal t ubules of the system arranged in tuft s. Each t uft opens by a short t ube

(duct ule of the fourt h order) on the ou t er side of t hetertiary ductu le. The terminal tubulesare long and usually cylindrical

,t hough some of them show bead - l ike swel lings at intervals.

The whole of the tubules borne on a secondary duct ule form , in t he nat ural posi t ion ofthe organs

,a wel l - marked secondary lobule

,and al l t he secondary lobules of a primary

ductule form a primary lobu le. These lobules,t hough t hey are placed at fairly regular

intervals along the bi le - duct,and roughly correspond on the two S ides of the body, are not

segmentally arranged in the abdomen. Moreover, owing partly to the fact that t he primaryductules run

,not straigh t upw ards round t he hind - gut, but also somewhat backwards, the

lobu les come t o overlap one another on t he dorsal side of the gut , and their arrangementappears a t first less regular t han it real ly is. Besides those arranged on the above syst em

,

each bile - duct bears , on i t s ant erior port ion , a cert ain number of t ubules direct ly sessi le uponi t. These gradually diminish in size from behind forwards, ti l l a short s t retch of the bi leduct

,j ust behind t he opening

,is left qu i t e free from them. H indwards the bi le - duct begins

by the j unction of two primary ductu les.

We may speak of the whole of the lobu les of each side as forming a righ t and aleft l iver respectively. The left liver

,t hen , is somewhat smal ler than the right, t hough t he

difference is not very marked. At the hind end of t heabdomen t he left l iver passes dorsalt o the right. The whole of the structures j ust described are bound together by s t randsof connective t issue carrying blood vessels. This circumst ance, combined w it h the soft , easi lybreakable t exture of t he liver, makes that organ rat her difli cult to unravel.

C. T he H ind - gu t . (Pl. III. fig. G .) B ehind the mid - gut,and stretching from it to

the anal opening in a direct course,runs t he chitin - l ined hind - gut. Its widt h is even for

thefirst three - fourths or so of its length and t hen it increases gradually t o a greater diameter.There are no rectal coeca.

Ou twardly,the surface of the hind - gut is smoot h

,but wit hin it is thrown into ridges.

These ridges,which are t welve in number

,are most marked at t he hind end , grow gradually

lower forwards,and finally fade away in the middle of t he first abdominal segment, some

distance from the end of the mid - gut. They are not all of equal size, being alt ernately largeand smal l or rather high and low

,for there is not much difference in breadt h. Each bears


a series of smal l t ransverse ridges , and each of t hese is again made up of a row of fouror five minute oval beads set with their long axes lengthways in the in t estine

,that is

transversely to the ridge which each row of them composes. The whole of t hese s t ruct uresare

,of course

,covered wit h a delicate chitinous pel licle. In t he anterior region of the

hind - gut,where t he ridges are wanting, the inner surface is covered wi t h similar smal l

beads,less regular in shape, set in longit udinal rows. The transition from the mid to the

hind - gut is hardly visible ou twardly in fresh specimens , but in a gut which has been fora few hours in spiri t i t is very easy to see , owing to the difference in cal ibre broughtabout by t he shrinkage of the mid -

gu t . Internally, the beaded appearance of t he hind - gutmakes t he dis t inct ion a sharp one. In the natural position of the organs in t he body agreat part of the hind -

gut is hidden by the overlapping lobules of the liver,but more

or less of the hinder part usually comes in t o sight amongst the lobules.

The anus (fig. 18) is sit uated on t he under side of t he telson. It is guarded by astrong valve in the form of an oval calcified convex plat e (1 , fig. attached to t he softventral wal l of the telson in fron t of the anus, and proj ecting back under it. If thisplate be parted from the body of t he telson , t he anus is seen as a longi t udinal sli t att he bo t tom of the hol low between i t and t he telson. B et ween these two s t ruct ures, t hen ,t here is a sort of cloaca, in t o which the anus opens. The aperture of t hecloaca is surroundedby a thick growth of hairs. The result of this arrangement is that t he opening for t he

discharge of fa eces is directed , not vent ralw ards , in which case it would be liable t o bepressed against t he columel la of t he shel l and t hus obstructed , but backwards , and isfurt her protec t ed from pressure by t he stou t plate beneath it. Possibly t he obj ect of thehairs round the opening is to preven t the en t rance of faecal matter as the animal shiftsin its shel l

,though it is remarkable t hat t he shel ls of t hese voracious creatures con t ain

quite a smal l quan t ity of dung. In the Gala theinea and B rachyura, groups W hose memberscarry the abdomen pressed more or less s t rongly against the underside of the thorax

,the

same arrangement is found in a less complet e form , the anus being usual ly prominent anddirected more or less backwards.

(ii i) T hev a scula r sy stem (fig.

The rather large,muscu lar hear t (19, fig. 19) l ies , as usual in t he thorax

,close under

the dorsal carapace behind the cervical groove , surrounded by the thin - wa l led pericardial sinus,

in which i t is suspended by the three pairs of fibrous alaecordis,and wi t h which it

communicat es by the usual t hree pairs of ostia. In shape it appears roughly four - sidedfrom above , the anterior end being drawn out into a low prominence from which arise thethree anteri or arteries. The hind end has also a low, rounded , median bulge. At abouta t hird of i t s length from t he hind end t here is on each side a notch. The dorsal surfaceis sligh t ly convex ; the sides slope inwards to the ventral surface, w hich is flat

,and is raised

,

at t he hind end , into a round , median lobe, from which t he sternal and abdominal art eriesarise.

Ar ter ies. Seven vessels are given off from the heart. At the front end there arisesa group of t hree—a median ophthalmi c and two lateral antennar ies. A short distance behindt hese t he paired hepa ti cs are given off from t he ventral surface of t he heart ; and at thehind end t he ventral lobe bears t w o median ar t eries, a more anterior, ventrally d irecteds terna l, and a more posterior, terminal abdomina l.G.

L. A. BORRADAILE.

The ophtha lmi c ar tery (1 , fig. 19) runs straight forward over t he roof of the stomachwithou t giving off any branches , and finally divides into two vessels which diverge forwardsat an acu t e angle, to supply the eyes. There is no swelling

,or median prolongation

beyond the bifurcation. The antennary a r ter ies (2, fig. 19) run forwards and outwards fromtheir origin on each side of the ophthalmic artery. On t he inner side each gives rise t o

an ant erior and a posterior gastric artery,which supply t he musculat ure of t he s t omach.

At about t he level of t he ant erior gast ric art ery there is given 05“ on t he ou ter side astou t branch , which runs downwards and forwards t o t he green gland. At a short distancein front of the post erior gast ric artery a less import ant out er branch arises

,to be dis t ri ~

bu t ed t o t he muscles of t he adj oining part of the t horax . In front of the st omach t he

dimini shed artery is con t inued forw ards to supply the antennae. B esides the above,t here are

o t her smal ler and more vari able branches.

The hepa t i c ar ter ies (5 , fig. The posi t ion of t he liver in the abdomen deprivesthese arteri es of t he function which gained them t heir name in o t her Decapods. It isperhaps best, however, to ret ain i t

,rather t han add to an already overladen t erminology. The

hepat ic art eries, then , in Coenobi ta do not supply t he liver, but run downw ards and forwards from the heart t o supply the sides of the s t omach. The sterna l a r tery (6 , fig. 19)on leaving the heart runs direct ly downwards

,passes on one side or other of t he in t estine

,

and ends by going t hrough the oval hole in the t horacic ganglion and dividing immediatelyunderneat h t he lat t er int o two branches

,bne of which passes direct ly forwards in the middle

line and is known as the ventral thoracic art ery,whi le the other runs directly backwards.

This latter is t he homologue of the vent ral abdominal artery of ot her Decapoda. It hashere very l it t le right to the name, but to avoid confusion , and also because i t does actually

F IG.19. Arter ial system of Coenobi ta

clypea tus . A. Arteries of fema le fromabove; B.Left or lower branch of ab

dominel artery of ma le (en larged). 1.

Oph thalmi c. 2.Antennary. 3.Anter iorga s t ri c. 4.Posterior gast ric. 5 .Hepat ic.6. S terna l. 7.Ventra l thoracic. 8.Ven

t ral abdomina l. 9.Dorsal abdomina l. 10.Righ t branch of abdominal. 11. Leftbranch of abdomina l. 12.Ovarian. 13.

Spermat ic. 14. B ranches to li ver. 15 .

Ovary. H ind-

gut. 17. Righ t bileduct. 18. Left bi le duct. 19. Hear t

(represen ted too large).



freely. The axis is at t ached at one point t o t he side of the thorax,and there the b

ent ers it. The lamel lae diminish in size towards each end of t he axis, whereby the wholegi ll becomes spindle - shaped. The gil l - formula is as fo llows 1

Anter ior Poster iorPodobranch

ar throbranch arthrobranch P leurobranch Tot al

rud. rud. rud.

The arrangements for t he supply of blood t o the gi l ls consist, as usual , of afferentbranchial vessels

,arising from t he sternal sinus, and of efferen t vessels leading to t he

peri cardial sinus and so t o t he heart .

The lining of the br anchi ostegi te is a thin , smoot h membrane. There is no spongytissue such as is found in Ocypode or vascu lar tuft s as in B irgus. According t o Bouvier z ,

blood enters the branchiost egi t e from t he large sinus which encloses the st omach , and leavesby the great vein

,which may easi ly be found a t the hinder edge of t he organ. This

start s in front as a smal l vessel and runs backward round t he lower edge of t he branchios tegi te, enlarging as i t goes. Finally it curves inwards with the binder edge of t he

branchiostegite and , when t his rejoins t he body, passes on t o the pericardium. I did notfind it possible by inj ect ing t his vessel to irrigat e any definite plexus in the branchiostegi te. The coloured fluid passes wit h great readiness int o t he space be tween t he st rongou t er and t he delicat e inner wal l of the organ

, but i t is here con t ained in a loose andirregu lar syst em of lacunae

,or rat her in a single cavity divided by st rands of t issue.

The gi ll chamber. Provi sion for moi s tening the gi lls. The t hird leg in i t s normalposit ion , pressed up against t he soft branchiost egit e , inden t s t he latter in such a way asto limi t t he branchial chamber to a comparat ively small region in t he binder and upperpart s of t he t horax. This chamber is widely open behind , so t hat t he gills can generallybe partly seen w it hout li ft ing t he branchiost egi t e. The animal seems t o be able to increase t his opening at wil l

,bu t over t he great er part of t heir surface t he branchiost egi t e

lies fair ly close above t he gills. There is thus no at t empt a t t he formation of anyt hinglike a lung. The free edge of t he branchiost egite is incurved , and the t rough t husformed is lined wi t h hairs and usually very moist . Anot her hairy surface , possibly of import ance in the reten t ion of water, is t o be found on the wall of the thorax above t he

gi lls,between t hem and t he origi n of the branchios tegite.

1 The gil l - formula of Pagurus i s the sameas that of 2 Bouvier , C.Bend. c x .pp.1211 fl. My observaCoenobi ta savethat t hegi lls on theth irdma x i ll iped and firs t t ions , w h ich confirmthoseofB ouvier ,weremadewhen I w asleg arebet ter developed. Eup agurus difiers in having on ly in ignorance of h is researches on t he blood supply of the

onepleurobranch—that on thefourth leg. branch ios tegite.

L AND CRUSTACEANS. 8 5

The source of the mois t ure is a problem of some obscurit y. It is of the very rarestoccurrence for a Coenobita to be found in t he w ater except at t he breeding period ‘.Ortmann states quite distinc t ly that , out of many hundreds, he has never seen one in thewater. Nor have I mysel f succeeded any bett er in this respect. That they do visit thesea

,however

,I am convinced by two facts. First

,that Mr St an ley G ardiner had t he

good fortune t o find a specimen of C. perla tus in t he water a few feet from the lagoonbeach in M inikoi. It is true that t his was a female

,but there were no eggs or young,

nor any t races of them , on

'

her abdominal l imbs. Secondly , t hat the body of a freshlycaught specimen is always moist

,and in a great many cases (certain ly t he majority of

C.per la tvs and C. rugosus ) t he shel l act ual ly contains a smal l quant ity of salt water?

Another possible method of moistening the gil ls must not be overlooked. In someexperimen t s to test the effect of drying t he gil l chamber it soon became evident thatthis was an impossibi l ity. By the careful use of cotton - wool and blott ing—paper, a greatpart of the moisture could be removed

, bu t absolutely dry i t was quite impossible tomake t he chamber. After a time t he limpid salt wat er became replaced by a more st ickyfluid

,which frot hed wi t h t he violen t effort s of t he scaphognat hit e. Of course i t is possible

t hat t his was blood , flowing from wounds in the del icate cu t icl e lining the chamber, madeduring the process of drying. But I was unable to det ect any such wounds , and thesame t hing happened in each of several experiments. Another explanat ion of the phenomenon is that t he fluid was provided by exudat ion through t he lining of t he chamber.No doubt in t his case the exudat ion was abnormal in quanti ty and qualit y. But it servedt o indicate a possible met hod of keeping the gil ls moist . In support of this i s t he factt hat specimens made approximately dry 3 w ith cot ton - wool and then placed in a dry woodenbox , l ived, and remained in good healt h for weeks. No doubt, i f t he suggested processof exudation takes place , t he different species of t hegenus are dependent on i t to differentextent s. One would expect, for instance, that i t would play a greater part in speciessuch as C'. sp inosus, which oft en live at some dist ance from the sea

, t han in C. r ugosus ,which is general ly t o be found close t o sea - water.

The movements of the scaphogna thi te. On raising the forepart of t he branchiost egi t eof a Coenobi ta the plat e on t he second maxi l la, to t hemovements of which the respiratorycurren t in water- living Decapods is due , wil l nearly always either be found to be inmot ion or short ly begin to move. When the animal is placed in water

,eit her fresh or

salt,i t is easy t o show , by means of a lit t le carmine or ot her coloured fluid

,that t he

ordinary current is produced here also. I t enters a t t he hind end of the branchial chamberand between t he last two legs , passes over t he gil ls and t hrough t henarrow passagewhichleads downwards and forwards from t he upper part of the gi l l - chamber proper

,and final ly

issues under the antenna. I have not been able t o observe a reversal of t he current , suchas that described by B ohn “ in many Decapods. The meaning of t he movement of thescaphognat hit e when the animal is on land is not clear. I t is possible t hat i t s object is

,

by act ing as a fan , to create a draught of air t hrough t he gil l chamber. In consideringthis view, however, we are met by the difficu l t y that there are oft en long pauses in t he

movemen t s, and that removal of the scaphognat hit e has no percept ible effect on t he animal.

1 Seebelow, p.91. ki ll theanimal.2 Seebelow

,p.91. 4 Bohn , 0.Bend.cx x x v.p.5 39.

3 I f thedry ing process becont inued too long, i t i s apt to

86 L. A. BORRADA I LE.

I t may be that the st imulus of li fting the branchiostegi te is sufficient to set t he appendagein mot ion , but in this case one would expect t he movemen t s eit her t o cease directly or tocon t inue as long as t he branchiostegi t e is raised. This is not what happens, the movement sbeing sometimes steady and sometimes fit ful and irregular. A third explanation, and t he

one which a t presen t seems the most probable , is that we have here an inst ance of avestigial habit , ret ained aft er it has ceased t o be of use t o t he animal. Las t ly, i f it couldbe shown that both sexes are in the habi t of going in t o the water a t frequent int ervals

,

yet another solu t ion of the question could be offered. For in that case i t would be possibleto suppose that t he mot ion of t he scaphognat hite was retained on accoun t of i t s beingindispensable t o the animal under water and at the same time

,for some physiological

reason,not susceptible t o inhibition for long periods and thus perforce cont inued on land.

But this view would require assumpt ions, which there is no j ustificat ion for making.

Abdomina l respi ra t i on. At the time of my sojourn in the Island of M inikoi , I wasunaware of Bouvier’s 1 researches on t his point . My ow n observat ions were much less com

plete t han his,but I can confirm his st at ement s wi t h regard to t he various channels

carrying blood back from t he abdominal wal ls to the pericardium ,at least as regards the

dorsal pair, the ventral I fai led t o observe. While he paid considerable at t ention to theanat omical side of t he question

, Bouvier does not appear to have made any experimentsto test his t heories. It i s interest ing to observe that if t hegills of bot h sides be cut off,leaving smal l stumps t o avoid loss of blood (i t would be bett er to ligature the gil ls in afu t ure experiment), t he animal is stil l capable of living. Indeed one, on which I performedthis experimen t , l ived several days , and finally escaped from t he vessel it was confined in.Taken in conjunc t ion w it h the fact t hat t he action of t he scaphognat hite may be suspendedwit hout harm to the animal

,this fact seems t o indicate t hat abdominal respiration is of

considerable import ance in Coenobi ta. I t is furt her interesting t o not e t hat the soft skinof t he abdomen is always damp. Possibly the obj ect of the hairs and fleshy processes ont he ventral surface of t he abdomen is as much to retain wat er as to play any part inrespiration by movement , as Bouvier suggests. I t would certainly appear

,from the elaborate

precau t ions t aken in various groups of land Decapoda to ensure t he presence of moist ureon t he breathing organs, as though respirati on were , in them at least , impossible exceptt hrough a moist surface.

v. K i dney s (g reen glands ) (Pl. III. fig. H).

The kidney of Coenobi ta is a large oval cushion , of a pale greenish colour in the livinganimal , placed in the head on each side of, and rather behind , the brain , and behind t he

base of t he an t ennae. The surface of the cushion is not even,but raised in to a number

of irregular rounded lobes,except in t he midd le of t he upper side, where a space is left

smooth , and forms a depression amongst the lobes. The hilum of t he gland is in frontand on t he ou t side. I am quit e unable to distinguish , by inj ection or ot herwise, any vesiclesuch as is found in nearly al l other Decapods and is especial ly well developed in t he

Pagurids. The only ot her instances in which this does not occur are quoted by Marchal

1 Bouvier , Bull. Soc. Phi loma th., Paris n. p. 194 with the per icardium on each s ide by two veins—a longB riefly put theappara t us consis t s of a tegumen tary dorsa l and a short ven tra l one. Thetwo veins of each s ide

plex us , fed from theabdomina l s inus and communicat ing join beforeentering thepericardium.


8 8 L. A. BORRADAI LE.

which t here is but one perforation—that for the s t ernal art ery ( 11 , fig. On each sideof t he ganglion arise

,a t wide intervals

,three st ou t nerves for the first three legs. In

front and behind , t he longi t udinal commissures j oin it as two more slender strands,one

close on each side of the middle line but qui te dist inct . Behind the nerve for t he t hirdleg of each side , between i t and the longitudinal commissures, t he small nerves for t he

fourth and fi ft h legs arise. In t his region t here are also several very fine,thread - like

nerves arising from t hedorsal surface of t he ganglion and running backw ards.

In front of t he nerve for the first leg, between i t and the c ircumoesophageal commissures

, half- a - dozen nerves for the mou t h part s arise (9,fig. These come off, not

as in t he thoracic ganglion of Pagnrns figured by Bouvier,at one level

,along t he edge of

the ganglion,but at different poin t s scat tered over i t s surface in that region. Two , in fact ,

arise qui t e on t he dorsal surface. H ere again there are several smal l nerve - threads intermingled w i th t he stouter nerves leaving the ganglion. The nerves of t he two sides arenot absolutely symmet rical but the asymmet ry is not very marked.

The abdominal cha i n. Leaving t he t horacic ganglion the longit udinal commissures runbackwards as far as t he beginning of the six t h abdominal segmen t . During the whole oftheir course they remain dist inct save where , in each segmen t , a ganglionic swel ling bindst hem together. Nerves are given off, not only from t he ganglia, but also from the intermediat e por t ions of t he commissures. The nerves of the two sides are not strict ly symmetrical.

The brain and c ircumoesophageal commissures are easily exposed by removing t he stomachfrom above, when they are found lying respect ively against t he ant erior and ven t ral body - wall.The hinder part of t he c ircumoesophageal commissures , w it h the t horacic ganglion and thenerves leaving i t

,is protected by a st out endosternal skelet on of the macrurous type, j oined

by t he usual rib - like bars wi t h t he endopleurae and epimera , which form the rest of thein t ernal ske leton. The abdominal chain lies along the ventral side of the abdomen under themuscle - band.

vii. T he genera t ive organ s (Pl. III. fig. K , a—c).

A. Thema le organs. The testes lie imbedded in the dorsal part of the liver andmore or less completely hidden by t he lobes of t hat organ. They are not symmetrical ,t he right being placed farther back t han t he left . I t is also rather the larger of the t w o.Each consist s of a much coiled

,sacculated tube compacted into a firm

,elongate mass, en

larged at the hind end. From t his enlarged part arises the vas deferens , a somewhatcomplicated st ruct ure

,divided in t o t hree regions the conducting tube, (2) the spiral,

(3) the glandular tube. The conduct ing tube is a short, irregularly twist ed duct arisingfrom t he inner , vent ral aspect of the test is at its hinder end. In calibre i t is smallerthan the glandular t ube

,but rat her larger than t he spiral. This latter is a fine

, semitransparent pipe arranged in one close coi l of abou t thirty turns. The glandular tube isstou t , of an opaque white colour, and t hrown into many irregular loops. It runs forw ardalong t he inner side of the test is

,becoming less convoluted as it goes, traverses the

anterior part of t he abdomen,and finally passes almost insensibly int o the ductus ejacula

torins—a simple , uncoi led t ube, slightly wider than t he vas deferens.


Comparing these struct ures w ith t he same parts in Pagur istes and Eupagurus‘,the

fol lowing points appear. The testis is more compac t in Coenobi ta t han in eit her of t he

Pagurid genera. In shape it represents, with its hinder swel ling, 3. more definit e form ofthe type foreshadowed by Eupagurus , which is, i n t his respect

,intermediat e between

Pagur i s tes and t he land genus. On t he ot her hand, in having one Spiral sect ion (inst eadof t w o

,sundered by a straight t ube), as also in the smal l calibre of this region , Coenobi ta

approaches Pagur i stes rather than Eup agurus. But in the shortness of the very narrowtube which comes immediately before the spiral

,that is which separat es the spiral and

t he conducting tube proper,i t again resembles Eupagurus. The ma le opening is in t he

usual posit ion on the coxopodite of the fifth leg of each side. The port ion of t he j ointwhich bears the opening is always more or less prominent and

,in some cases (0. r ugosus ,

C. per la tus) is , on the right side of the body, prolonged into a penis - l ike process. InC.perla tus , however, in which t his feature is very marked

,bringing about a complet e dis

simi larit y between the openings of the tw o sides, there is no sign of degeneration in theleft testis , and i t s duct may contain sperm. The first abdominal segmen t being

,in both

sexes , unprovided with appendages, there is here no copulatory organ of the type commonin other Decapods.

B. Thefemale organs. The ovar ies are a pair of simple cylindrical structures lyingin t he abdomen one on each s ide of the median line close above t he hind -

gu t along themiddle portion of its course. In C. clypea tus t he ovaries are separat e. In C. rugosus andC. perla tus they j oin for a short distance at t heir hind ends ? The oviduc t is a simpletube

,arising from the anterior end of t he ovary and running straight forwards to i t s small

round opening on the ventral face of the coxopodite of t he third leg.

O. The rela t ive numbers of thesex es seem fairly equal , a collection taken at haphazardgiving somet imes a smal l preponderance of one

,somet imes of the other.

viii. Rep roduc t i on. W he t her Coenobi ta have a definite breeding season or not , and,if so, when i t occurs are quest ions t hat sti l l remain to be settled. Certainly, females witheggs may be t aken t hroughout t he summer in Ceylon 3

. The copulat ion is anot her pointdeserving furt her investigation. Very li t t le is known on this subj ect as regards mostDecapods. Paguri stes is said t o insert a penial process into the vu lva of t he female“, andpossibly t he same use is made of t he “ penis ” of C. per la tus. That of C

'. rugosus is t oo

broad for insert ion,but is no doubt of use in placing the sperm in some required position.

At the same time the reproductive organs of the left side of the body show every signof being funct ional in these

,as in the o t her species of t he genus

,and t he problem is thus

complicat ed by t he probabi li t y that their sperm is deposi ted in a different way from t hatof the right side of the body.

The eggs are carried in large masses attached to the long hair s on the wel l - developedl imbs on the right side of the second

, t hird and fourt h abdominal segmen t s of the female.They are arranged irregularly along the hairs 5 and fixed

,as in ot her Decapods , by an

1 G robben,“ Bei t.Kennt.Mann.Geschlech t sorg.Dekapoden

,

” Arb.Zool.I ns t.W ien,x.2

2 In Pagurus they aresa id to rema in separa te.3 Mr S tanley Gardiner sugges t s to me tha t thesou thwes t monsoon i s thema in breeding season in theMa ldives.Hew as in theseis lands dur ing thenorth -east monsoon , and

G.

found grea t diffi cul ty in finding females wi th eggs a t tha t

season. Thesamei s thecasei n Cey lon.4 Ortmann, B ronn

’eThierrei ch, v.2, p.1075 .5 In B i rgus they are in clumps a t in terva ls a long the

ha ir. Borrada ile, W il ley’s Zool. Resul t s, ” Pt.v. p. 5 8 5“ On theY oung of theRobber Crab."

L.A. BORRADAILE.

ou t er shell,prolonged into a stalk which adheres to the hair. The ripe egg is el lipsoidal

,

and meas ures,when preserved in spiri t , 7 5 mm. by 7 mm. in C. perla tus , and in C. r ugosus

rat her less. I t is thus a little smaller t han t hat of Birgus‘. In t wo species at least

(0. r ugosus and C. per la tus) the young are hatched as zoaea larvae 2 of a type much te

sembling those of other Pagurineas and washed off into t he sea. They are not swept out

in the respirat ory current in t he manner described by Bate ‘ for E upagurus nor is i tnecessary for t he female t o issue from her shel l, t hough she may do so if kept underwater t oo long in a st ate of capt ivit y. I t i s possible

,though perhaps not very likely

,t hat

the habi t s of some of the other species , especial ly th ose which live a t great er distancesfrom the water

,may have led to t heir young being hatched in a lat er stage ”, as is said

to be t hecase wit h some land crabs“. A zoaea which seemed identical wit h t hat of C.r ugosuswas taken at night with the t ow - net in the lagoon at M in ikoi.

The embryonic sk in,which encloses t heDecapod zoaea before hatching

,and is ret ained

for a varying period in different forms , is here lost very shortly aft er leaving the egg. I t

is of simple form and much resembles t hat found by Sars in ot her Pagurine zoaeas, t houghI have not been able t o discover any feathering on the processes of the glove - l ike s t ruct u rewhich encloses t he telson and i t s spines. The larva (fig. 21) is rat her smal ler t han that

of B irgus ’ (the lengt h is, in 0. per la tus , 3 mm.and in 0. rugosus 25 mm., as against 35 mm.in B i rgus ), w hich it otherw ise very closely re

sembles in al l but a few point s of detai l. Thecarap a cehas a curved , point ed rostrum of moderat elength. The hinder edge is hol lowed , and a t eachsidea rounded side - lobe replaces the spine usuallyfound here. The “

abdomen consist s of five segments and an end part, which shows indicat ionsof i t s coming division to form the six t h segmen tand t elson s. The telson (a ,

fig.21) has the usualfan shape with a rat her deep median not ch. On

each side of t henot ch are five bris t l es , increasingin lengt h from t he first to t he fourt h , which islonger t han t hefift h. These brist les are feat hered.The ou t er angle is provided wi t h a st ou t t oot h.None of t he ot her abdominal segmen t s are armed ,save t hefi ft h

,which has a single st rong spine on

each side a t t hehind end. There areeight pairsof limbs

,including a rudimentary pair of third maxi l lipeds. These limbs are almost exact ly

like t hose of t he first zoaea stage of B i rgusg

,ex cept for the mandible (b, fig. which has

F IG.21. Zoaea of Coenobi ta perla tus. a.Telson en larged.b. Mandi bleenlarged.

1 Borrada ile, loc.c i t.2 Borrada ile, P.Z.S ., 1899, p.937.3 G.O. Sars , Arch.M a t h. og Na turv id., x n r. p. 133‘1 Ba te, quoted by S tebbing “ Crust acea , p.164, London ,

1893.

5 C.clypea tus a few mm.long arefound in thejungleinM in ikoi in sea shells. For C.diogenes , seebelow, p.91.

6 Gecarc inus sp. Westwood, Phi l.Trans., 1835.7 B orradaile, W il ley’

s Zool.Result s ,” lac.c i t.3 Theseindicat ions arenot a l together wan t ing in B i rgus ,

so far as can beseen in theil l - preserved ma ter ia l from the

Pacific a t my disposa l.9 For figures of these see Borradai le, W i lley’

s“ Zool.

Result s ,” 100.c i t.



ot her Decapods in being pract ically omnivorous 1—belonging, in fact,to the class of

scavengers. They pick up various dead sea - animals along the beach,and get o t her refuse

elsewhere. On occasion , even cannibalism is not beyond t hem , when one of their numberis wounded or ki l led. They are able t o find t heir food in the dark

,and t here can be

litt le doubt t hat this is due t o a sense of smell , by which , rather t han by sight,they are

most likely guided in t heir search. Substances wit h a strong odour such as roasted coconutor t he fruit of the Pandanus (which has a distinct and characteristic smel l) are part icularly attractive to t hem. I t is interesting to watch t he way in which a Coenobi ta wil lpause in i t s walk , unfold its long ant ennules2 (see p. 70) and seem t o explore the airwit h t hem

,waving their flagel la gent ly t o and fro over i t s head. The same limbs are

advanced and held over the food during feeding. They find the fruit of the Pandanus ont he ground , but wil l also climb the bushes t o reach it. On one occasion an individualwas found in a Hernandea pelta ta abou t twenty feet from the ground

, but its obj ect ingoing there is hard to surmise. The curved end - j oin t s of the legs can clasp tigh t ly quitesmal l twigs

,and t he sharp claws w it h which they are provided are, no doubt, of use in

climbing. The food is often dragged for some distance , and in the case of fruit s this is,no doubt , a met hod of distributing the seeds, which should be t aken accoun t of inconsidering t he extension of vegetat ion over the land surface of a coral is land.

In t he matter of habi t at ions, t he choice is as varied as in that of food. Every avai lable kind of land or sea gast ropod shel l is used

,provided that i t be of t he right size and

not so encumbered with spines as t o be awkward in use. Weigh t seems to be of litt leaccount, the heavy Turbo - shel ls being especial favourites wit h C. clypea tus , al t hough onlycomparat ively small individuals can use t hem. One small specimen of C. r ugosus in Ceylonwas carrying the empty tube of a Serpulid worm. C’. spinosus is known t o use the nutshellof Ca lophyllum inophyllum

s, and C. clypea tus , when it has grown too large for gastropod

shel ls , takes the half of a coconut“. A case even occurred in which a broken glass t ube wasmade use of 5.

Ortmann 6 has point ed ou t t he e x istence in C. r ugosus ( t races of the same st ruct ureare found in most 0. per la tus) of an apparatus , which he regarded as adapted for produc ing a sound. This consist s of a row of smal l ridges on the ou t side of t he great ( left )chela and a longi t udinal ridge on the underside of t he second walking leg of t he sameside. I t is int eresting to know t hat a number of t hese creat ures shut up in a large t in

1 Ortmann [Bronn’

s Thierreich , r. 2, p. 1234] seems tobeof theopin ion that Coenobi ta i s purely veget arian in i t sfeeding, and quotes Dah l and S treet s in support . But

Catesby [Na t H i s t.Carolina s , n . p. 33] long ago observedtha t C.di ogenes will occas iona l ly t akean ima l food.

Theseappendages are, of course, thosein wh ich a sense

resembling tha t of smel l ha s been found to beloca ted in other

crust aceans. Lack of t imeun fortunately preventedmefrommaking any ex periment s on thi s poin t , wh ich I am surewould repay inves t iga t ion i n Coenobi t a.

3 B orrada ile, P.Z.S., 1898 , p. 459. Gast ropod shel lsareprobably scarcewhereth is i s done.

4 In such cases as t h is,themeans by whi ch thecrea t ure

reta in s i t s houseareof in teres t. Theabdomen i s doubledforwards under thet horax , so as t o present theroughenedsurfa ces on the6 th abdomina l l imbs to theins ideof the

shell , and a t thesamet imethe4 th and 5 th thoracic l imbsmakeuseof thes imi lar pa t ches on their propodi tes. Evenso thehold on theshel l i s but a feebleone(seep.

5 A s imi lar casei s men t ioned by B rock [quoted by Or tmann, B roun’

s Thierrei ch, v.2, p. The in teres t ofth is observat ion l ies in i t s bearing on thequest ion of the

means by wh ich t he an ima l recogn ises an objec t as beingsuit ablefor i t s “ house.” In th is case i t seems far more

l ikely tha t the senseof touch w as employed t han tha t of

s igh t. On theother hand, an individua l removed from i t s

shel l , will makefor another shel l placed a t somedis tanceina manner wh ich seems to indicatetha t th is i s seen.

6 Ortmann , B ronn’s Thierrei ch, v.2, p.1249. H ilgendor f

i s sa id to havefirst ca lled a t tent ion to theex is tenceof t h isarrangemen t.

LAND CRUSTACEANS.

box,gave out con t inually a low , chirping sound, though it was not possible t o discover

how they did this ’. The object of t he sound is not clear,and t heories on t his subject

must wait ti l l the quest ion of the hearing of t he creatures is decided. They certainly oft enseem to be affect ed by sounds , but whether this may not be due t o other vibrat ionsst arted at the same t ime in the eart h and surrounding object s is sti l l doubtful to me

,my

informat ion on t hepoin t being limited to d esul tory observat ions made at M inikoi before I wasobliged to leave the island.

While they are not st rictly noct urnal animals,the land hermit- crabs are cert ainly more

active by night than by day. They seem to avoid the heat of the sun by preference,and

to shel ter, during t he middle of the day , in nooks and crannies. Their habit of crowdingtogether in any place

,where food is to be found

,makes t hem seem gregarious—which

,in

a stri ct sense of t he w ord they probably are not. At the same time t hey may sometimes be found col lect ed in considerable numbers for no apparent cause.

They are not given t o fighting t o t he same extent as t he hermi t - crabs of t he sea.When molested they wit hdraw quickly in t o t heir shel ls

,closing the opening with t he big

left chela, which is special ly adapt ed for this purpose. In this condition a Coenobi ta in ast rong shell , such as that of Turbo argyrostomus must be an exceedingly tough nut for mostanimals t o crack , and can also fall from a considerable height without injury. If they areunable to wi t hdraw into t heir shel ls they wil l sometimes endeavour t o defend themselveswit h the same powerful l imb that is used to close the shel l , but i t oft en requires a considerable amoun t of teasing t o induce th em t o do this. The grip of t he great chela isexceedingly st rong and will easily snap a twig which t he animal has been made to seize.

If an individual be placed on i t s back , t hat is wi t h t he mouth of the shel l uppermost

,i t wil l thrust out its body t i l l t he shel l overbalances and t he animal is able to

recover the normal posit ion. In moving , t hey crawl obliquely forward and t o the left . Fromt he accounts of eye - witnesses 2

,i t would seem t hat O’. di ogenes is considerably swifter t han the

Indo - Pacific species.

A series of ex periments carr ied out with t he obj ect of inves t igating t he vit alit y of theanimals in fresh and sal t water led to t he fol low ing conclusions

(i) Continuous submersion under water is always fatal after a more or less prolongedperiod. (i i) There is great individual variat ion in the length of this period , bu t t he sexesdo not differ great ly ? (ii i) The creat ures are very sensitive to t he effects of overcrowding.

(iv) Tinned vessels aremore injurious than enamel led ones. (v) There is a difference be tweenthe powers of resistance to fresh and sal t water in favour of the latt er, but t he difference isnot so great as might have been expect ed. The greatest length of li fe in salt water reachedby any individual during t heexperiment s was 6 days, but t his was except ional. In fresh wat er24 hours was not exceeded. (vi ) Of the t hree species , C.per la tus showed t hegreatest vit alityin w at er and C. clypea tus the leas t , but t henumber of individuals of t he lat t er experimen t edwith was small. (vii) B efore death t he animal general ly

,but not invariably, comes out of

1 B roderip [Zool.Journ., i v.p.205 ] quot ing from theold 3 Fema les wi th eggs do not differ from others in th isEncyclopedic (Paris , 175 1) sta tes tha t C.diogenes makes respect , and thehat ch ing larvaedied a lmos t as soon in sa lt

a sma ll sound when i t is sei zed. as in fresh water. Had i t been poss ibleto rig up a dipper ,2 Soane, Na t.H i st.Jama i ca , 11.p. 272 ; Ca tesby , Na t. no doubt someof thesemigh t havebeen reared.

H is t.Caroli nas , 11.p.33.


its shell. (viii ) An individual , seemingly dead, may oft en be revived by being placed inthe air , j ust as a marine crab, dying from ex posure to t he air, will come to life in wat er‘.

Some 200 specimens , most ly of the species C. rugosus and C. perla tus , were ex perimented with. The fresh water used was rain gat hered in a “ galvanised iron tank. Owingto t he limit ed number of vessels avai lable and lack of t ime

, t he experiments were not ascomplete as they shou ld have been. Stil l , t hey may perhaps serve to induce some morefortunat ely situat ed observer t o carry on the enquiry. To obtain good resu l t s

, each individualshould be placed by itse lf in a vessel wit h a considerable quan t ity of wat er

,which shou ld

be frequently changed. It would be in t eresting t o t ry and induce the animals to feedunder water. The few at tempts, which I made at this , were unsuccessful. That deat h isnot due t o starvation is , however, certain from a considerat ion of t he fact t hat t hey wil ll ive for weeks wit hou t food

,if not placed under water.

3. SOME NOTES ON THE LAND CRABS or THE GENUS Ocypode.i. Ocypodecera tophtha lma (Pallas).

This species lives in burrows in t he sandy strand of the lagoon. Large warrens of t heseburrows extend along t he shore j ust below extreme high - water mark. They are thuscovered and dest royed at high tide

,but during many hours are completely exposed. The

mou t h of t he burrow is not always situat ed so as t o be covered by an average tide, but

the lower par t alw ays fal ls in a t high wat er,owing t o t he loosening of the sand. The

crab , which await s this event at t he bot t om of i t s hole , is thus buried for some hours.When the t ide falls i t works its way out and repairs i t s burrow

,and until the wat er

retu rns may be seen moving about t he shore near t he opening2,and running wi t h ex tra

ordinary sw iftness when threat ened wit h capt ure. Under these circumstances the first impulseof t he animal is to make for i t s hole

,bu t , if cut off from this , i t will seek safet y in t he

wat er. The crabs double readily,and a favourite device wit h t hem is t o remain st i l l ti l l

the pursuer approaches and t hen suddenly dart off. These manoeuvres are ca rried out onthe t ips of t heir slender legs and wit h t he eyest alks raised so as to survey as wide a fieldas possible , and give the impression t hat the animals possess in t el ligence of a highorder. When final ly seized

,t hey make vigorous efforts to defend t hemselves with t heir

chelae.

As might be supposed from t heir burrows being submerged a t high tide , and fromtheir readiness to take refuge in the sea

,t he crabs are able t o endure immersion in salt

wat er for a number of hours,though they are eventually ki lled by it. Fresh wat er, on

the o ther hand,is rapidly fatal

,two hours being the longest period that any of t hem sur

vi ved an ex perimen t .

The burrows are inhabi t ed each by a single crab ,and Maj or Alcock 3 has shown t hat

the st ridulating apparatus,possessed by this species in common w ith mos t others of t he

1 In th is connect ion i t i s interest ing to note tha t the s t imula t ion t o givea few beat s.heart of the common hermit - crab [Eup agurus bernhardus 2 In thehot test part of theday , when thesun shines ful l

and theshorecrab [Ca rc in ides moenas may upon thebeach , they areless ac t ivethan at other t imes.often befound to bebeat ing somet imea fter every outward 3 Alcock , Ann.M ag.N.H. x . For 0.cera ts ign of l ifei s lost from ex posure to thea ir. Even when i t op h tha lma seeAnderson.Journ.As.S oc.Benga l, Lv.has s topped, i t may somet imes be induced by mechanica l



of a few hours. The Ocypodes appear to find their food by sight,rat her than by smel l

l ike Coenobi ta.

ii. Ocyp odecordimana ,Desm.

While it is alive,this crab is easily dist inguished from 0. cera tophtha lma by its darker

and more brownish colour, but when preserved in spiri t it takes on much the same dul lgreyish - green hue as the other species. The two are

,however

,always quite easily separated

by t he absence, from 0. cordimana ,of t he stridulat ing apparatus found in 0. cera tophthalma

as in al l t he rest of t he genus. Their habits are also considerably different. Unlike thestrand - haun t ing cer a tophtha lma , cordimana l ives inland , digging its burrows in the lightsandy soi l along the pa t hs and open spaces of the island. Instead of being directed downwards , these burrows usually take t he form of more or less horizon t al galleries wi t h two

,

or somet imes t hree,openings. I have not found leaves, seaweed or food of any sort in

those that I have opened,but they run among t he roots of the vege t ation and t hese may

perhaps serve for food.

Two poin t s of in t erest are raised by t he fact s just mentioned. In the firs t place itis worth remarking t hat t he darker colour of 0. cordimana harmonises bet ter wit h t hatof i t s eart hy environment than would the sandy hues of 0. cera tophthalma. In t he second

,

the ex istence of a species w hose burrows are sit uat ed on land,wel l above t he t ide - mark

,

invalidat es the conclusions,as to t he raising of t he land in Diego Garcia, drawn by Bourne 1

from t he presence of Ocypod- holes in cert ain si t uat ions t here. The form of these holeswould have to be careful ly investigat ed before any such conclusions could be drawn fromthem.

III. A LIST or LAND AND FRESH WATER CRUSTACEANS COLLECTED IN THE MALDIVEISLANDS.

I am indebted to Mr S t anley G ardiner for the notes incorporated in the following list.

B RACHYURA, CATOM ETOPA.

Fami ly O cy podidae. G enus Ocyp ode, Fabr., 1798.

1. Ocypodecera tophthalma (Pal las), 1772.

For references, see above , p. 67.G enerally distribut ed t hroughout the group.

2. Ocyp odecordimana Desm., 1825 .

For references see above, p. 67.Al l the larger islands of the group except in Suvadiva and Addu at ol ls.

G enus Uca ,Leach , 1815.

3. Uca annulipes (H.M. 1837.

Gelas imus annulipes , H.M.- Edwards. Crust . I I.p.5 5 , pl.XVIII.figs. 10—13 Alcock,

Soc. B engal LXIX. ii. 3 ,p. 35 3 (1900) [references].

Uca annulipes , Ortmann , Zool. J ahrb. Syst . X. p. 35 5Mangrove Swamp

,Furnardu,

M iladumadulu at oll.1 Bourne, P.R.S., vol.43, p.445


Family G rap s i dae. G enus Geograp sus , S t imps., 18 5 8.

4. Geograpsus grayi (H. M. 18 53.

For references see above, p. 67.In every inhabited island of the group.

G enus Metasesarma ,H.M.- Edw ., 18 53.

5. Metasesarma rousseauai H.M.- Edw ., 185 3.

For references see above, p. 68.G eneral distribution throughout the group in damp land , especially at dry edges of

mangroveswamps. Not found in Addu at ol l.

Family G eoc a rc i n idae. G enus Cardiosoma , Latr., 1825.

6. Cardiosoma carnifex (Hbst ) 1794.Cancer carnifex ,

H erbst,

“ Krabben I I. v. p. 263, Pl XLI. figs. 1 , 2Cardiosoma carnifex , Alcock , As. Soc. B engal , LX IX. 11. 3 , p 445 (1900) [references].

Of general dist ribu tion t hroughou t the northern atolls. Especially common in Miladumadulu and Mahlos. G eneral ly makes its burrows under coconut trees at the edges ofkuli or swamps , the openings being oft en covered at high tide. Not found in Addu atol l.

ANOMALA,PAGUR I NEA.

Fami ly Coenobi t i dae. G enus Coenobi ta ,Latr., 1826.

The distribution of this genus in the Maldives is somewhat pecul iar. In nearly everyisland one species at least is found, but it is rare to find three. Two are frequent ly metwith in the same island , and C. rugosus is t he most common. In Goidu al l four are metwith

,and H u lu le

,Male

,has per la tus , r ugosus and clypea tus.

7. Coenobi ta perla tus H.M.- Edw .,

For references see above , p. 68.Of general distribution throughou t the group.

8. Coenobi ta rugosus H.M.- Edw., 1837.

For references see above,p. 68.

Of general dist ribution throughou t the group.

9. Coenobi ta comp ressa -

s. G uerin , 1830.Coenobi ta comp ressa Guerin , Voy. Coqui l le

,I I.2

,p. 29

Coenobi ta compressa s , Ortmann , Zool. J ahrb. Syst. VI. p. 318 (1892) [references].Goidu

,and probably elsewhere.

10.

I

Coenobi ta clypea tus Latr., 1826.

For references see above,p. 68.

Of general dist ribu tion throughout the group.G.

L. A. BORRADAI LE.

CAR I DEA.

Family Palaemon i dae. G enus Leander , Desm.

11. Leander debi li s (Dana), 185 2.

Pa laema n debi li s , Dana, U.S. Ex pl. Ex pd., Crust. I. p. 5 85 , PI.XXXVIII. figs.6, 7Leander debi lis, Ortmann , Zool. J ahrb. V. Syst. p. 5 1 5 (1890) in part.

Ortmann (loc. ci t.) included under t his species , as varieties , several ot her Leandersdescribed by various authors. Among these was L. longi carpus, S t imps , 1860, which hassince been identified by de Man and Goutiere with L. concinnus Dana, 18 52. The presentcollection contains examples of a form al lied both t o L.debi lis and to L. concinnus , thoughmore closely to the former t han to t he latt er. In fact t he point s of difference between thepreserved Specimens would certain ly be sufficiently smal l t o j ustify their inclusion in a singlespecies were it not that they exhibit considerable differences in colour and habit at. For noteson these I am indebted to Mr G ardiner.

The present species has 6—9 t ee t h on t he underside of t he rostrum and 4—6 above,

and the carpopodit e of t he second leg j ust reaches t he end of t heantennal scale. Its size israt her great er t han t hat of L. gardiner i. When alive i t is co lourless , bu t has eggs ofa bril liant dark green.

The only localit y in which i t was found in the Maldives was a kuli surrounded bya mangrove swamp in Landu, M iladumadulu atol l. It was here few in numbers andsolit ary in its habi t s.

12. Leander ga rdiner i , n. Sp.

Under the rost rum of this Species are 4—6 t eet h,and above i t 5—6. The second

leg of ful l grown individuals is longer than in L. debi li s and i t s carpopodit e exceeds t heant ennal scale. The size is less t han t hat of L. debilis , t he largest specimen being 33 mm.in length. The third flagel lum of the antennule resembles t hat of L. debi lis , and not L. con

cinna s , in being free for less t han half i t s lengt h. The colour is intermediate between st rawand brown , t he eggs being of a darker shade of t he same ; t hebranchios tegi tes a si lvery - white.

This prawn was found in enormous numbers at the edges of a large fresh - water ku li inEkasdu

,M iladumadulu at ol l. Both it and the former species are t hus of interest in t hat ,

contrary t o the usual habits of t hegenus, t hey live in fresh wat er. The allied L. conci nnusl ives indifferently in fresh

,brackish or salt wat er.

I SOPODA, ON I SCO I DEA.

Family L i g i i dae. G enus Ligia , Fahr., 1798.

13. Ligia ex ot i ca Roux , 1828.Ligia ex ot i ca , Rou x , Crust. Médi t. I II. Pl.XIII. fig. 3

Fairly common t hroughout the group on boat s, ships , wharves , et c. At Mahugudu,

M iladumadulu atoll , extremely common al l round the shores of t he island on the rocks.The specimens from the latter locality are somewhat smaller t han those from other placesin t he group or t han those met wit h in Ceylon or M inikoi , and their colour is difi

'

erent ,

consist ing of a whit e ground covered with microscopic dark spot s, giving a greyish appear



G enus Phi losci a ,Latr. 1803.

16. ? Phi loscia graci lis , Budde - Lund , 1879, var.

The col lection contains specimens of a Phi losci a taken in several different localit ies,

but unfortunately al l more or less damaged. They are a t al l events al l ied to P. graci lis

(Budde - Lund Prosp. 2, 1879, and Isop. Terrest, p. 220, but differ from the descript ion given by the author of that species in the fol lowing poin t s

1. The inner ramus of t he uropod , though wel l - developed, is shorter than the outer.

2. The fifth trunk segment is no darker in colour than t herest .

Common in damp vegetable matter throughout the group, from Mahlos to Addu atol l.

17. Phi losci a ,sp.

Damaged specimens of a Phi losci a from Hedufuri , Mahlos at ol l.

Family Armadi lli di i dae. G enus Cubari s, B randt, 1833.

18. Cubari s murinus , B randt, 1833.

For references see above, p. 68.Addu and probably also other atolls.With the exception of the new species

,al l the above crustaceans are Indo - Pac ific in

distribu t ion. The only feature of interest in this respect exhibit ed by the collection isthe absence from Addu of several species found in the more northern atolls.


n.)

7000.

Pla teV.

NE M E RTE ANS .

B Y R. c. PUNNETT, B.A. (calms ).

Wit h Plates iv. and v.

To j udge from our previous knowledge of the Nemerteans of the Indian and EasternPacific Oceans, the collection of these worms made by Mr Stan ley G ardiner is, on t he

whole,much what one wou ld have been led to expect. The genera Eupolia and Drepano

phorus are wel l represent ed , as also is the family of t he Lineidae. The curious genusPr osadenoporus, first met with at Amboina

,is recorded for t he second time only. The

col lection also includes examples of two of the most widely distributed Nemerteans in thispart of the world

, i .e. Eupoli a hemp r i chi , and Li nens a lbovi t ta tus. The col lection is howevernotewort hy in that it adds to the fauna of these seas the first Mesonemer tean recorded.Cephalothr i a

'

a liena belongs to a genus , which has hitherto been recorded from the NorthAtlantic and from t he North - west Pac ific 1 only. It is however not unlikely that, as theNemertean fauna of the vari ous regions of t he globe becomes better known , this genusmay turn out to be almost cosmopolitan. Almos t al l the species are excel lently preserved,and I wou ld t ake this opportunity of thanking Mr G ardiner for the energy and troubleexpended in forming this col lection. In t he fol lowing pages a bri ef diagnosis of each newform is given

,and this is fol lowed by a more complete description. At the end of t he

paper I have appended a brief synopsis of the genus Drep anophorus.

SYSTEMATIC LIST.

I. PROTONEMERTINI z. (None)

II. MESONEMERTINL

Fam. Ceph a lot h ri c idae.

1. Cep halothr i x a liena, n. sp. Comparatively short, stout form with thick epithelium

and basement membrane. Mou t h beneath brain. No eyes. p. 102

1 Coe, W .R., Proc.W ash.Acad.S c., 1901 , p.19. Monograp h, 1895.2 The class ificat ion i s that given by Burger, Nap les

R . U.

III. HETERONEMERTINI.

Fam. E up oli i dae.

2. Eupoli a.hemp r z’

chi. (Ehrenberg, p. 1043. Eup olz

'

a i ndi ca , n. sp. Ex ternal ly marked by a single median dorsal and vent ralbrown line. Head much flatt ened. Circular head furrow marked by small longit udinal grooves inside. Ex cretory system connected with oesophagus p. 104

4. Eupolz’

a uni stm'

a ta ,n. sp. With single median black line. G reatly resembles

E .melanogmmma as regards in t ernal st ructure p. 106

Fam. L i neidae.

5. Cerebra tulus ma ldi vensis , n.sp. Small and somewhat flatt ened. Cutis feeb le. Vascularhead loop. Head slits reach pas t cerebral organ. No eyes. Frontal organ present.Excretory system short and wi t h one pair of ducts p. 106

6. Cerebra tulus macula tus , n. sp. Medium sized and somewhat flat t ened. Covered wi t hblack pigment spot s. Cu t is weak. Vascular head loop. Head slit s not pastcerebral organ. Eyes. No frontal organ. Excret ory system much attenuated andwith one pair of duct s p. 107

7. Cerebra tulus gardinen’

, n. sp. Medium sized and somewhat flattened. Cutis feeble.Vascular head loop. Head slit s prolonged j ust pas t cerebral organ. Eyes. Frontalorgan present. Excretory system wit h one pai r of ducts (occasionally more) p. 108

8. Cerebra tulus is churus , n. sp. Large and powerful ly buil t form. Cutis feeble.Vascular lacunar network in snou t. Head slit s end abruptly where ci liatedcanal comes off. N0 eyes. Small front al organ p. 110

9. Linens a lbovi t ta tus. Burger, 1890 p. 111

METANEMERTINI.Fam. Amp h i pori dae.

10. Prosadenoporus buerger i , n. sp. No external markings. Head glands just past brain.Cerebral organ smal l and w ith ventral opening. B lind gut with two pockets only.19 proboscis nerves p. 111

11. Drepanophorus roseus , n. sp. Much flattened. Cerebral organ has lateral opening.It does not extend behind brain. Eyes very numerous. Proboscis nerves22 p. 112

12. Drepanophorus car inas (Burger, 1890) p. 1 13

ACCOUNT OF THE SPECIES.

1. Cepha lothm'

x a liena ,n. sp. (Pl. IV. figs. 15 ,

Represented by 5 cc. of fragments abou t 15 mm. in breadth. Colour dirty white.Ant. end rounded and much contracted. No external markings to be dist inguished. Posteriorportion missing.

Locali ty. Felida atoll,Maldive Is. Dredged from 20 fathoms.


104 R. c. PUNNETT.

The most marked point of difference between C. a liena and the rest of the genuslies in the relat ive posi t ions of mouth and brain. In al l t he other species whose anatomyhas been described the mouth lies behind the brain

,and usual ly some dist ance behind.

The struct ure of the epi t hel ium with its double layer of nuclei is more complex than int he rest of t he genus. In t hepresence of an inner circular muscle layer 0. a liena resemblesC. signa ta and differs from the other species of Cephalothr iw. The great t hickness of the

basemen t membrane also is unusual. Moreover t o j udge by the figures in Burger’s Monograph(Taf. II.) t he rhynchocoelom is less spacious and the blood vessels considerably more so t hanis usually the case.

2. Eupoli a hemp ri chi (Ehrenberg, (PI. IV. figs. 2,

Syn. Nemer tes hempr i chi , Ehrenberg, SymbolaePhys icae, 1831.Eupoli a brocki , Burger, Zeit . f. wiss. Zool., 1890.E up oli a bempri chi , Burger, Naples M onograph , 1895.Eup olia hemp r ichi , Punnet t, Wil ley

’s Zool.Results, Pt . V., 1900.Loca li ty. Found sparsely on every reef.

Several specimens of t his common and widely - distributed species were obtain ed from thereefs both in the Maldive and t heLaccadive Islands.

The size aft er preservat ion varied from 65 cm. to 104 cm. in length. The last - mentionedspecimen is larger than any ot her yet recorded. The fol lowing not e was made on the livingworm by Mr S tanley G ardiner : “ Lengt h when obt ained out of a hole in the rock of the reeffla t 5 ft. 6 in. 1676 B lack st reak down cent re of back and same t ransverse behindhead.” These markings become brown aft er pre servation. There was in t his specimen also apigment fleck on t he t ip of t he head and furt her a ventral line of about the same widt has t he dorsal ex tending along the whole length to the col lar j ust behind the mouth. Here

,

however, i t is much t hinned (Pl. IV. fig. and the collar is also incomplete ventral ly. Inmarkings as wel l as in size this specimen is int ermediat e between the figures which B ii rgergi ves of E. brocki (loc. ci t. 1890,

Taf. I. fig. 10) and of E.mediolinca ta (Zool. Jahr. Ed. 7 ,Abt. f. Syst. Taf. VIII. fig. The thickness of the dorsal and ventral lines and also ofthe t ransverse one is exceedingly variable. In one case t he line ' may be at least one - thirdof t he body - wid th , whilst in anot her i t is exceedingly fine. Sometimes again the dorsalstripe is the broader of the t w o

,whilst in others t he ven t ral exceeds the dorsal in wi d t h.

Such considerat ions render it not unlikely that the species described by Burger asE . mediolinea la may be merely a local variety of E. hemp r i chi. The species has a widerange, extending from New B rit ain in the East Pacific

,across the Indian Ocean to the

Red Sea.In a previous paper I ment ioned that the excret ory sys t em in a specimen of thi s

species (Punnet t , 1900, loc. ci t.) was not provided wi t h ducts. This led me to make sectionsright through t his region in a specimen from the Maldives. The result was a confirmationof t he previous st atemen t that t he excret ory system is here unprovided with ducts openingto t he e x terior.

3. Eupolia indica ,n. sp. (Pl. IV. figs. 8 , 12, 13 ,

Locali ty. Hulule,Male atoll

,Maldive Is. From reef.

NEM ERTEANS. 105

A single complete specimen was obtained measuring 22 cm. in length, and with an

average bread t h of 35 mm. In external appearance much wrinkled and somewhat flattened.There is a wel l - marked circular head furrow. On the dorsal surface there is a very palebrown stripe about 1 of the body - width. A similar st ripe occurs also on t he vent ralsurface.

The skin is fashioned in the style characteristic for the genus. The closely packedlayer of cutis glands is abou t three times the thickness of the epithelium. The connect ivetissue layer of t he cutis i s rather thicker t han the glandular layer.

The circular muscle layer in the oesophageal region is rather thicker than the innerlongitudinal layer. The outer longitudinal layer is stouter than the other two together.

The vascular system in the snou t region forms a distinct cephalic loop, though t he vesselsare very wide and separat ed only by a t hin parti t ion. This loop lies dorsal to the rhynchodaeum,

i .e. in the same position as the anastomosing lacunae of ot her species of Eup olia.In ot her respects the vascular system shews no peculiari t ies marking it off from t he usualarrangement such as that found in E. cur ta.

The alimen t ary canal shews no marked pecu liari t ies.

The proboscis is peculiar in possessing only one layer of muscle fibres, i .e. a longitudinal

layer. The epithelium is comparatively low.

The excret ory system start s abou t 4 mm. from the tip of t he snout. It lies entirely ator below the level of the line joining the nervous side stems. I t extends over about 5 mm.It possesses no duct s to t he ex t erior nor any t races of such structures. Fine cords of cel lshowever

,which have almost the appearance of delicate ducts compressed to obscure the

lumen,pierce the glandu lar layer of the oesophagus (Pl. IV. fig. 14) and may be traced

t o the oesophageal epit hel ium. Such an arrangement recal ls t he peculiar condition foundin Eupolia melanogramma

’

, though the communicat ions differ probably from this lat ter speciesin not being functional.

The brain is much flatt ened in shape (Pl. IV. fig. its bread t h being about three timesas great as its depth. The dorsal commissure is t hick , straight , and short. The dorsal ganglionlies more lat eral than dorsal to t he ven t ral ganglion.

The cerebral organ is not large and is considerably flat t ened. The ci liated canal passesstraight to t he inner border of the organ where it dilates into a large spherical vesic lew ith high epithelium containing a few gland cel ls (Pl. IV. fig. A smal l quan t ity ofglandular secretion also occurs wi t hin t his expansion of the ciliated canal. The dorsal lobeof the dorsal ganglion reaches some way over the cerebral organ. The opening of the cerebralorgan is lateral and slightly ventral (Pl. IV. fig. The ci liated canal in this position opensinto the circular head furrow

,and t he circular head furrow in t his species is peculiar in being

marked by a number of smal l longitudinal furrows which extend entirely round it (Pl. IV.fig. The epithelium of these smal l head furrows is characterised by an absence of epithelialglands and by the numerous closely packed oval nuclei (Pl. IV. fig. Along the summitsof t hese small ridges occur smal l refractive bodies which do not stain. Such head furrowsare peculiar to t his species of Eapolia so far as is known, and recal l t he smal l head furrowsfound in a simi lar posi t ion in the genus Drepanophorus (

“ Kopfgriibchen

1 Punnet t , ! uar t.J ourn.M i c.Se., vol.44, p.116.

106 R. o. PUNNETT.Numerous smal l eyes occur on t he dorsal surface of the snout. They lie for t he most

part in the cutis,t hough some are found in the outer longitudinal muscle layer.

The head glands are w el l developed and are of t he t ypical Eupolia t ype. They reachbackwards as far as the level where t he excretory tubu les st art. Anteriorly they open int ot hewell - marked front al organ.

4. Eupolia uni str i a ta ,n. sp. (Pl. IV. fig.

Loca li ty. Hu lule,Male atoll


Two specimens were obtained. The larger measures 5 5 cm. in length,whilst the average

bread t h is 6 mm., and depth 4 mm. The smal ler specimen measures 26 cm. in lengt h and5 mm. in breadt h. In life the worms were white wit h a black stripe down the back. Thew idth of t he st ripe is less than 1 mm. In the larger specimen t he stripe stops short atthe head furrow (Pl.IV. fig. whilst in the smaller one i t is cont inued on to th e snout. Nomarkings on the ventral surface.

In ext ernal form t his species bears a great resemblance to E .melanogramma (Punne t t ,! uart. Journ. Mic. S c., vol. 44, p. This resemblance is found also in the in t ernalstructure. I t wil l be suffi cient here to indicate the main poin ts of difference.

The connecti ve tissue layer of t he cu t is is not so thick as the glandular layer. InE.melanogr amma these propor t ions are reversed.

The excretory system possesses numerous openings to the ex t erior,but it lacks t he

peculiar duct s opening into t he oesophag us which characterise E.melanogramma.

The arrangement of t he vascular syst em in t he cerebral r egi on is that characterist ic ofthe genus. In E.melanogram-ma the arrangemen t is slight ly different (loo. ci t. p.

With these exceptions t he anatomy of t he t w o species is pract ically identical. Asin E. melanogramma there is present also in this species a well - marked sub - anal nervouscommissure.

5 . Cerebr a tulus ma ldi 'uensi s , n. sp. (Pl. IV. fig. 5 , and PI. V. figs. 3 , 6 ,

Locali ty. M i ladumadulu atol l , Maldive Is. Dredged in 22 fat homs.

This Species is represent ed by about 30 mm. of fragments. The average breadt h isabout 17 mm. The more posterior port ion is much flat tened , not measuring more t han“3 mm. in depth. The anterior end is more rounded. The hind end is missing. Theproboscis pore is terminal

,and the mou t h is si t uated about 2 mm. behind t he tip of

the snout,which is somewhat point ed. Colour whitish aft er preservation. No markings

present.

The epi t helium is t hin and cont ains a few unicellular glands which stain deeply wit hpicric acid. The basement membrane is ex ceedingly t hin (Pl. V. fig. Just below i t t hereis a well - marked layer of circular muscle fibri ls and j ust beneat h t hese again a few longit udinal fibri ls which are separated from t he ou t er longitudinal muscle layer by a very t hinlayer of cutis glands. The whole cu t is is but feebly developed.

The ou t er longit udinal muscle layer in t he oesophageal region is about double t he

thickness of t hecircular layer. The inner longitudinal layer in t his region is of approximat elythe same t hickness as the circular layer. All t hese t hree layers are somewhat massive


108 R. C. PUNNETT

The epithelium is somewhat low, and t he nuclei form a well - marked layer a t t he

bases of the cells. In the outer clear portion of the cells are a few unicel lu lar glands,

which take a vi vid yel low stain with picric acid. A very t hin basement membrane separat esthe epi t helium from an exceedingly del icate layer of circular muscle fibrils. Below t heseagain come a few longi tudinal muscle fibrils

,whi ch are separat ed from the outer longit udinal

muscle layer by a layer of nuclei , apparen t ly representing t he cu t is glands. The pigment ,which gives rise t o the black markings of the worm

,is situated partly in t hecutis (Pl. V.

fig. 1) and partly in the nuclear layer of the epithelium. The outer portion of t he ou t erlayer of longitudinal muscles is more or less divided into bundles by strands of connectivetissue.

Of the muscle layers of the body wal l the circular muscle layer is rat her t hicker t hant he inner longitudinal but less t han hal f the thickness of the ou t er longitudinal layer.There are wel l - marked hori zont al fibres over t he mouth. The precerebral region is moreclosely packed wit h muscle fibres than is usually the case. There is no diagonal musclelayer.

The mou t h commences short ly before t he l evel where the cerebral organ terminates.There is no ventral gutt er to the in t est ine.

The vascular system shews a well - marked cephalic head loop. The buccal commissureis wel l developed. Short ly aft er t he terminat ion of t he excretory tubules the oesophageallacunae are gathered toget her into a single small lateral lacuna which extends to t he in

test inal region. H ere i t acquires a coat of parenchyma cells,and gives off dorsal commissural

branches , becom ing t he lateral blood vessel.

The proboscis is fairly well developed,and shews the three muscle layers characteristic

of the Lineid proboscis. Muscle crosses are present .

The excretory syst em commences soon aft er the hinder t ermination of the mont h. Thereis a median longit udinal tubu le running along a t t he level of t he nervous st em. Int o itsmal ler tubu les open a t intervals

,extending a lit t le way above and below the level of the

side st ems. They are however bot h smal l and few in number, t he whole excretory systembeing much att enuated. There is a single duct (Pl. V. fig. 10) on ei t her side about themiddle. The lumen of t hese ducts is also reduced , and i t is probable that t hey were notfunctional in t he living animal.

No gonads are developed.

The brain is not very large. The dorsal ganglia lie directly over t he ventral whichare closely apposed. B ot h dorsal and ventral commissures are ex t remely short. The mediandorsal nerve is well marked. Neurochord cel ls are present in t he vent ral ganglia t hough t heyare not so large as usual. The cerebral organ is fairly wel l developed though the gland cel lsin connec t ion wit h i t are somewhat scan ty. I t lies for t he most part dorsal to and justinside t he nervous side stems. The head slit s reach nearly t o t he brain and ext end backwards t o t he level where t he cerebral organ t erminates.

Numerous small eyes are present on the tip of the head just dorsa l and ventral tothe head slit s. Frontal organ and head glands are both absent.

7. Cerebra tulus yardiner i , n. sp. (Pl. IV. figs. 6 , 9, and Pl. V. 2, 4 ,Locali ty. M inikoi , Laccadive ls. From lagoon sand.

NEM ERTEANS. 109

Fragmen t s of t wo specimens preserved , i .e. about 73 mm. of one worm and 98 mm. ofa somewhat smal ler specimen. The lat t er specimen is almos t complete. The breadt h is fairlyuniform , being abou t 2 5 mm. in the larger and 2 mm. in t hesmaller specimen. The intest inalregion is somewhat flat t ened , t hough not greatly so. The head is sharply point ed

,t he head

slit s are long, and t he mou t h is long and slit - l ike. Colour when alive white or whi t ishflesh.

The epithelium is somewha t low and contains a large number of un icel lu lar oral glands.It rest s on a thick basement membrane (Pl. V. fig. 2) beneath which is a well - marked layerof circular muscle fibri ls. B etween t his layer and t he ou t er longit udinal muscle layer is thecu t is, which cont ains a few smal l bundles of longitudinal muscle fibres

,t races of cu tis glands

and a little connect ive t issue.

The muscles of t he body wal l are fairly wel l developed. The outer longi t udinal layer isabout double t hethickness of t he circular layer (Pl. IV. fig. 9) in t he oesophageal region , andt his again is twice as t hick as the inner longitudinal. layer. The oesophagus is provided wit ha separate and very delicate coat of longitudinal muscles. There is a strong layer of horizontalmuscles over the mou t h.

The mou t h commences a t the level where the cerebral organ t erminates,or sligh t ly in

front of t hat level. The in t estine possesses a well - marked ventral gu t ter. The hinder part ofthe oesophagus is much constrict ed.

In t he vascu lar system t here is present a cephalic loop. Shortly behind the t erminationof t he excretory t ubules the oesophageal lacunae are gathered into a single lacuna

,which lat er

becomes the lat eral blood vessel of the intestinal region as in 0.maldivensi s (p. 0.ma cu

la tus (p. The dorsal blood vessel leaves the proboscis sheath at the level where theoesophageal lacunae are gathered into a single lacuna

, i .e. rather more than 1 mm. beh indt he termination of t he excret ory tubules. The arrangemen t of the blood lacunae in the cerebral region conforms to the typical Lineid t ype.

The proboscis is long and fairly st out,being about 1 mm. in diamet er. I t presents the

usual three muscle layers and possesses t w o muscl e crosses.

The excretory t ubules first appear in sect ion j ust before the post erior limit of the mouth.They ex t end a litt le w ay dorsal ly and ventral ly to t he level of the nervous st ems. In onespecimen t here was a single duct on ei t her side, making an angle of about 45

°

(Pl. IV.fig.9,card.) wit h the nervous side stems. In the ot her specimen there was a single duct placedrat her in front of the midd le of the sys t em on the right side, whilst on the left side therewere four duc t s

,t hemost anterior of which had almost disappeared.

No gonads were found in eit her specimen.

The brain is large and the dorsal ganglion is more t han double t he size of the ventral.The median dorsal nerve is not wel l marked. Neurochord cel ls occur on the inner side ofeach vent ral ganglion short ly before i t s separation from t he dorsal ganglion. The mediandorsal nerve is not wel l marked. The cerebral organ is large (Pl. V. fig. 4 c), and the glandcells in connection wit h it are somewhat scanty. The head sli t s are deep

,reaching almost to

the brain. At the same time they are exceedingly wide. They continue deep aft er the ciliat edcanal has been given off

,and terminate a t a level slightly behind t heposterior end of the

cerebral organ.

1 10 R. o. PUNNETT.

A fron tal organ is present. Numerous smal l eyes occur,chiefly on the dorsal surface

and j ust beneat h the basement membrane. Head glands not well marked.

8. Cerebra tulus isobaras,n. sp. (Pl. V. fig.

Loca li ty. Hulu le,Male at ol l


Fragments of a single specimen amounting to 21 cm. in all. The oesophageal regionis rounded in shape, and the bread t h here is about 7 mm. The in t estinal region is somewhat flattened , and shews a wel l - marked median ventral longi t udinal groove. I t measuresabout 14 mm. in wid t h and about 5 mm. in dept h. No record has been preserved of t hecolour in life. After preservation it is a uniform pale dirty brow n.

The epithel ium is thin,and cont ains a few unicellular gland cells. The basement

membrane is very t hin,and beneath it is a fine layer of circular muscle fibri ls and then

some longit udinal fibrils separat ing t he last named from t he cutis glands. The last are

not well developed. The st ruct ure of the body wall bears’

a great resemblance to that ofC. haddoni (Punne t t, Proc. Zool. Soc. 1900, PI. LV. fig. from which i t differs in t he somewhat greater developmen t of t he cu t is glands.

The muscles of t he body wall are s t rongly developed. The ou t er longit udinal layer isabout half as thick again as t he circular layer. This in i t s turn is about double t he

thickness of the inner longitudinal layer. There is a thick layer of longi t udinal musclesbetw een t he proboscis sheath and oesophagus. It is con t inued round t he oesophagus as adelicate investmen t of longitudinal muscle fibrils.

The mou t h commences at t he hind level of the dorsal ganglion and lies below t he

cerebral organ.

The vascular system in the snout shews a system of lacunar anast omosing spaces inl ieu of a w el l - marked head loop. The hinder part of the cerebral organ is surrounded bya blood lacuna. The oesophagea l lacunae are spacious.

The proboscis is slender. I t is two layer as regards musculat ure, and shews twomuscle crosses. The wal l of t he proboscis sheat h is remarkably muscular. Its cir cularmuscle layer is j ust over half t he t hickness of t he circular muscle layer of t he body wal l.There is also a delicate ou ter investment of longi t udinal muscle fibrils cont inuous w ith thelayers separat ing t he proboscis sheat h from the oesophagus and circular musculature of thebody wal l respect ively.

The ex cretory system commences some way behind the mouth. The tubules extendven t ral to t he nervous side stems and also dorsally to theproboscis sheath. In the anteriorfragment out there was but one excret ory duct. Judging from other forms , i t is exceedinglyprobable that had more excret ory duct s been presen t some w ould have occurred in thecomparat ively large number of sect ions t hrough the ex cretory region. Consequent ly it isprobable t hat C. i soba r a s is a form wi t h but one pair of excretory duct s.

The brain is smal l compared with the bulk of t he animal. The fibrous core of thedorsal ganglion is about t wice the size of t hat of t heventral ganglion. The dorsal nervouscommissure gives off a median dorsal

,which for a very short distance behind t hebrain lies

in the outer longitudinal muscle layer,and not between i t and the circu lar layer as is


1 12 R. c. PUNNETT.

Two very large eyes are present on either side.As in t he other species of the genus t he head glands are exceedingly well developed

and reach back a lit t le way past t he brain. They open into a deep pit lined by modifiedect oderm resembling that figured by B ii rger (Naples Monograph , Taf. 18 , fig.

The only account of the genus Prosadenopor us is that given by Burger in foundingthe genus on several species from Amboina (Zeit. f.Wiss. Zool. 1890, p. I have t herefore thought i t fi t t ing to name this new species in honour of the zoologist t o whom ourent ire previous knowledge of the genus is due. From the account given in this publicationi t will be seen that the present species differs from the forms t here described chiefly int he absence of external markings, the number of proboscis nerves , and in the ventralposition of t he opening of the canal of t he cerebral organ.

11. Drep anophorns r osens , n.sp. (Pl. IV. fig.

Loca li ties. Hulule , Maldive Is. From boulder zone. Na ifaro, Fadiffolu atol l , Maldive Is.

Several specimens,most of which weresomewhat fragmentary. In three complete specimens

the length varied from 50—64 mm. and the breadth from 5 mm. aft er preservation.The posterior and middle portions of t he body were exceedingly flatt ened. Colour in lifepink to red.

The epit helium is h igh , and contains numbers of elongat ed unicel lu lar glands wit h smal lgranu les in side (

“ S tabchenzellen It is separated from the circu lar muscle layer by athick basement membrane. The relative thickness of the t hree layers , epithel ium ,

basementmembrane, and

“ circular muscle layer, is as 4 : The longit udinal muscu lar layer iswel l marked

,being of about the same t hickness as the preceding three layers together. A

thin diagonal layer occurs between t he circular and longi t udinal.

Mou t h and proboscis pore open separately into a common depression on the ventralsurface near t he anterior ex tremit y. The lat t er opening is the more anteriorly placed. Atfirst wit hout glands the oesophagus begins to shew large deeply st aining unicel lular glandsin the cerebral region. There is a well - marked ventral blind gut passing forward from the

j unction of oesophagus and in t est ine. I t does not however reach forw ards quit e two - thirdsof t he distance bet w een i t s point of origin and the brain. The intest ine is a smal l roundtube whose area in cross - sect ion is not much more t han double that of one of the nervousstems at the same level.

The vascular sy st em conforms t o t he usual metanemer tean arrangement. It containsnumbers of large oval corpuscles which render i t exceedingly conspicuous in sect ion. Thelateral vessels l ie just dorsal t o the nervous stems, and t he commissural vessels which theygive off pass direct ly up between the intestinal divert icu la to t he median dorsal vessel.Ot hers, however, pass right round the ends of the intestinal diverticula before reach ing themedian dorsal vesse l.

The proboscis sheath presents t he usual basket - like arrangement of its muscle fibreswhich characterises t he genus. The divert icula are long and slender, and pass round t he

int est inal pouches to reach as far as t he nervous st em. The proboscis shews the usual

NEM ERTEANS. 1 13

metanemertean s t ructure, and cont ains twen ty - two nerves. The length of the proboscis isabout half that of the body.

The excret ory tubules are fi rst found in sec t ion j ust after the binder end of t he cerebralorgan. When first seen in sec t ion the tubules lie laterally and dorsally in the gelat inousconnect i ve tissue just beneath the longitudinal muscle layer. As it issues from the nervousring the lat eral blood vessel on ei t her side makes a short bend to the lat eral edge of t hebody before taking up i t s position j ust dorsal to the nervous st em. The port ion of thevessel forming this bend is somewhat dilated

,and the excretory tubules enter into close

relat ion with i t. The excretory pore is situated about ‘

5 mm. behind the terminat ion oft he brain on eit her side. The duct in each case passes round the outer side of thenervous stem and opens on the ventral surface directly beneat h it. The excretory tubu lesextend backwards nearly to the intestinal region

,being all the way closely apposed to the

lateral blood vessel.

G onads undeveloped in all the specimens.

The brain is la rge and round , and the dorsal ganglia have abou t 4—5 t imes the area oft he ventral in section. The dorsal commissure is very long and slender. The distancebetween t he nervous st ems in t he intest inal region is about half t he w id t h of the body.A smal l median dorsal nerve occurs be t w een t he circular muscles and t he basement membrane. Just behind the dorsal commissure a smal l but wel l - defined nerve leaves the dorsalganglion on either side to run along the side of the proboscis sheath j ust above thediverticula of that st ructure. I t passes into the longit udinal muscles of this region. I t isnot to be dist inguished behind the oesophageal region. W hether this nerve supplies the musclefibres or whether it passes through the muscles and basement membrane t o pass to the skinis doubtful. The fact that it comes off t he dorsal ganglion probably indicates a sensorynerve. The nervous stems post eriorly form a wel l - marked supra- anal commissure. Onepairof neurochord cel ls is present in t hebrain.

The cerebral organ is large,being in cross section about 2, of the area of the dorsal

ganglion at its largest. It lies exactly lateral to the dorsal ganglion and terminat esposteriorly at the same level. It possesses a cap of gland cel ls over i t s hinder portion ,which extends rather furt her forwards dorsal ly than ven t rally. The opening of the cerebralorgan is si t uated rat her in front of the organ and exactly on the lateral edge of the body.There arewel l - marked li t tle longit udinal head furrows extending some waydorsal ly and ventrally. The cerebral organ is supplied by one large nerve which comes off

the hind end of t hedorsal ganglion._

Numerous large eyes are present arranged in four longit udinal rows (cf.Pl. IV. fig.There are well - marked head glands

, but t hese are short, not extending more than one - third ofthe dist ance from t he t ip of the snout to the brain.

12. Drep anophorns cerinns . (BURGER. (PI. IV.figs. 10,

Li tera ture. Burger, O. Zeit. fur Wiss. Zool., 50 Ed. 1 890.

Locali ty. Suvadiva atoll , Maldive Is. From 38 fathoms.

The account given by B ii rger is exceedingly short . The smal l size and the peculiarst ruct ure of the cerebral organ , however, render t he species fairly easy to iden t ify. I haveadded below a somewhat ful ler account than that given by this author.G.

1 14 R. C. PUNNETT.

Two specimens were obtained measuring abou t 25 mm. in lengt h and 5 mm. in breadt h.The specimens were not quite so flattened dorso - ventrally as most members of t his genus

,

owing t o t he fact that t heir bodies were bu lged out by ripe gonads. The proboscis isst out and measures about 10 mm. in lengt h. The colour is a very pale pinkish - brown aft erpreservat ion , and no external markings are distinguishable.

The epithelium is thick and rests upon a very thin basement membrane,which is not

nearly so st rong as the circular muscle layer j ust beneath i t. The longitudinal musclelayer is st rong.

The vascu lar system present s the usual met anemertean arrangemen t .

The mou t h opens jus t behind the proboscis pore, but the separat ion betw een the t w oopenings is complet e. There is not even a depression into which bot h open as in mos tspecies of the genus. In the cerebral region t he greater part of the oesophagus (Pl. IV.fig. 10) lies squeezed bet ween the brain lobes. The blind gut extends abou t half- way fromi t s point of origin towards the brain.

The proboscis sheat h shews the usual basket - like arrangement of the muscle fibres. Thefirst few rhynchocoelomi c divert icu la are exceedingly smal l, and even in the posterior oeso

phageal region t hey reach barely as far as the nervous stems. The proboscis shews t he usualst ru cture. I t cont ains but 14 nerves.

The excretory tubules reach forward nearly as far as t he opening of t he cerebral organ.The excretory pore (which is paired) is about t he same distance behind the brain as thefront end of the brain is from the tip of the snout. I t is ventral in posi t ion

,and in

the specimen cut t he duct is much reduced and was probably not functional.

The gonidial sacs of the specimen cu t contained large ripe ova. The sacs st art int he post erior oesophageal region

,and then open j ust dorsal to the nervous stems on ei t her

side. In t he intest inal region t he openings are ven t ral near the lat eral edge of t he body.

The brain is very large,the dorsal ganglion being about 6—7 times t he size of the

ven t ral (cf. Pl. IV. fig. The t w o lobes of t he brain are closely apposed , both dorsal andvent ral commissures being comparat ively short . There i s a small median dorsal nervebetween the basement membrane and circular muscle layer. The dist ance between t he t w o

lat eral stems in t he in test inal and in t he oesophageal regions is more t han half t he totalbread t h of t he body of t he worm. There is a wel l - marked supra - anal nerve commissure.

The cerebral organ is very large. I t is si t uated entirely behind t hebrain. I t s openingis rat her behind t he level where i t first appears in sect ion‘. I t is lateral and sligh t lyven t ral (of. P1. IV. fig. The struct ure of t he organ has already been described byBurger in the paper referred t o above. The essential peculiarit y of t he organ is t hegreat lyelongat ed glandular tube which st ret ches backwards above the nervous st ems for a dist ancewhich is rat her more than double t hat of t he organ itself. The head furrows are wellmarked and extend round nearly two - thirds of t he circumference of the head region.

Large eyes are found,though t he number of t hese is rat her smal ler t han in most

species of the genus. About 26 in all are present.

1 B urger gives a reconst ruct ion of theorgan (Zei t.fur rela t ion may bea local variat ion and does not appear suffiw i ss.Zool. 1890, p.240) from w h ich i t would appear t hat cient to separatethepresent form as a dist inct species.theorgan i s not placed en t irely beh ind thebra in. Th is


1 16 R. o. PUNNETT.

Pos it ion ofcerebra l organ Gon idia] pores Peculi ar fea t ures

> 30 19—20 1 latera l t oeach intes t ina ldivert i culum

Latera l Bes ideand > 30 24

rat her beyondbra in

Beh ind bra in 2 vent ra l to Anteri or rhynea ch intest inal chocoelomicdivert i culum divert i cula fuse.Nephridia l pore

at pos t.end of

nephridi umD. lat us (7 ) Latera l, ratherdorsa l

D. cerinus and Latera l and Beh ind bra in 14 1 vent ro - la tera l Very long post h is paper ra ther vent ral t o ea ch terior glandularintest ina l process to

divert iculum cerebra l organ

1 lateral t oea ch Long posteriorintes t ina l glandulardivert i culum process t ocerebra l organD.boreal is Vent ral Bes ideand not 3 dorsa l t o Anterior rhyn

beh ind brain ea ch intest inal chocoelomicdivert i cul um divert i cula fuse

and formexpans ions over

bra inD. roseus Th is paper Lateral Bes ideand not > 30 22

beh ind bra in

List of more importan t or morerecent papers referring to thegenus(1 ) Hubrech t , A.A.W . Thegenera of European Nemerteans crit ica l ly revised. Notes. Leyden Mus.

(2) H ubrech t , A.A.W . TheNemerteans , Cha llenger Repor ts, Vol. 19. 1887.

(3) Joubin , L. Archie deZool.E x p.et Gen. 1890.

(4) B ii rger , O. Zei t.fii r w i ss.Zool. 1890.

(5 ) Joubin , L. Fannefranga i se. 1894.

(6) Burger, 0. DieNemert inen. Fauna and Flora des Golfes non Neapel. 1895.

(7) S t aub , J . Semon’s ZoologischeForschungsreisen. 1900.

(8 ) Punnet t , R.C. W illey ’s Zoologi ca l Results , Pt.V. 1900.

(9) Punnet t , R.C. Proc. Zool. Soc. 1901.

NEMERTEANS. 1 17

EXPLANATION OF PLATES.

basement membrane.cut is.c ilia ted cana l.ex panded end of c i l ia ted cana l.cerebra l organ.dorsa l blood vessel.dorsa l lobeof dorsa l gangli on.dorsa l gangl ion.epi thel ium.ex cret ory duct .ex cret ory t ubule.front a l organ.glands of cerebra l organ.head furrow.head s l i t.la tera l vascula r lacuna.mouth.c ircu la r musc le layer.

mcc circu la r muscle layer of cut is.

PLATE IV.Cerebra tnlns macula tus. Dorsa l and vent ra l v iew of anter ior end. x 2.

Enpoli a hemp r iehi. Dorsa l and vent ra l V iew of anteri or end. x l.

E. hemp r i chi. Anter ior end of anot her specimen from dorsa l surface. x 2.

E. n/ni s tr i a ta. Anter ior end from dorsa l surface. x 1.

Cerebr a tn ln s ma ldivensi s. Vent ra l View of anteri or end. x 4.

C. ga rdiner i . Vent ra l v iew of anter ior end. x 2.

Drepanophorns roseu s. Vent ra l v iew of anterior ex t remi ty. From spec imen clea red in cedarOil. x 5 .

Enpoli a i ndi ca . Sect ion t hrough region of head furrow , shewing out l ineof bra in. x 22.

Cerebr a tulus gardineri . Sect ion t hrough oesophagea l region, shewing rela t ive t h ickness of

musc le layers and pos i t ion of ex cret ory du c t . x 15.

Drep anophorn s cer inus. Sect ion t hrough bra in region immedia tely after the sepa ra t ion of

thedorsa l and vent ra l gangl ia. x 45.

D. cer inn s. Sec t ion just anter ior t o t he open ing of t he cerebra l organ. x 45.

Enpoli a i ndi ca. Sect ion t hrough cerebra l organ and h inder pa rt of dorsa l gangli on. x 45.

E. i ndica. G rooved epi t hel ium of c ircu la r head furrow as seen when morehigh ly magnifiedthan in Fig. 8.E. indie-a. Sect ion t hrough port ion of oesophagus in ex cret ory region, shewing a cord of

cells from an ex cret ory t ubu le" pierc ing t he oesophagea l glandular layer and reach ing t o

t he inner surfaceof t he a l iment ary cana l. x 80.

Theasterisk marks the level wherethedorsal vessel emerges from theproboscis shea th.

interna l c ircula r musclelayer.longi t udina l muscle layer.longit udina l muscle layer of cut i s.interna l longi t udina l musc le layer.outer longi t udina l muscle layer.median dorsa l nerve.la tera l nervous s tem.

eyes.

oesophagea l epit hel ium.oesophagus.oesophagea l glandula r layer.oesophagea l vascula r la cunae.probosc is.p igment .probosc is shea th.rhynchocoelom.

vascular cepha l ic lacuna.vent ra l gangl ion.

1 18 R. o. PUNNETT.FI G. 15 . Cephalothr ix aliena. Sec t ion t hrough body wa ll in hi nder part of oesophagea l region.

Thegrea ter port ion of t he sec t ion is somewha t dorsa l to the level of t henervous stem.x 80.

FI G. 16. Cepha lothri x a liena. Sect ion through anter ior part of oesophagea l region. x 45.

PLATE V.FIG. 1. Cerebra tulus macula tus. Sect ion t hrough sk in of oesophagea l region. x 160.

FIG. 2. Cerebra tu lus ga/rdiner i. Sect ion t hrough sk in in oesophagea l region. x 110.

FIG. 3. Cerebra tulus ma ldinensis. Sect ion t hrough sk in of oesophagea l regi on. x 110.

F IGS. 4—7. Sec t ions t aken a t interva ls t hrough the h ind bra in and cerebra l organ of t he followingspeciesFIG. 4 (a—c). C.ga/rdiner i. A t interva ls of 80 p..

FIG. 5 (a—c). C.macula tus. At interva ls of 100 p.

FIG. 6 (a—c). 0.

F IG. 7 (a —c ). C. i schn rns. 5 0 p. bet ween a and b ; 150 p. bet ween 6 and c.

ma ldinensis. A t in terva ls of 80 F'

F IGS. 8—10. Schema t ic reconst ruc t ions t hrough t he anterior end of severa l Lineidae. Th-e buccalva scu la r vessels

,t he oesophagea l nerves , t he probosc is and i t s shea t h

,have, for t he sake

of simpl ici ty, been om i t ted. The oesophagea l vascular lacunaehave a lso been representeda s s implevessels.FI G. 8. 0.gamdiner i. x 10.

FIG. 9. C.ma ldi vensis. x 10.

FIG. 10. C.ma cula tu s. x 10.

Date post:	20-Feb-2023
Category:	Documents
Upload:	khangminh22
View:	0 times
Download:	0 times

The Fauna and Éeographfy of the M5 Dive and Lacca'dive ...

Documents