AJi. J. Ecol. 1984, Volume 22, pages 1-22 Nocturnal ecology of the springhare, Pedetes capensis, in Botswana T. M. BUTYNSKI* Department of Wildlife, National Parks and Tourism, Gaborone, Botswana Summary The springhare, Pedetes capensis, was studied in the Kalahari Desert, Republic of Botswana, from August 197 1 to August 1974 inclusive. Night censuses totalling 253 h provided data on springhare habitat selection, social grouping, activity pattern and the effect of moonlight and weather on activity. Utilization of dry lakebeds (pans) by springhares was 2.3-3.4 times greater than on the surrounding habitats. Springhares were strictly nocturnal. The population, as a whole, exhibited one activity peak every 24 h. Group size and distance from burrows were directly related to time after sunset. Low temperatures and moderate to heavy precipitation both reduced springhare activity, while moonlight reduced the distance to which springhares moved out onto pans. Except for adult males, all sex/age classes of springhares participated in group formation to the same extent. Apparently little social cohesion existed within groups. Territoriality was not observed but may occur in the area immediately around the burrow. Re‘sume‘ Le lievre sauteur (Pedetes capensis) fut etudit dans le desert du Kalahari, au Botswana, de aoGt 1971 a aoGt 1974. Un total de 253 heures de recensements nocturnes a fourni des donnees sur le choix de I’habitat chez le lievre sauteur, son organisation sociale, ses types d’activites et l’effet du clair de lune et du temps sur celles-ci. L’utilisation du fond des lacs esskchtis (pans) par ces animaux est de 2.3 a 3.4 fois plus important que les habitats avoisinants. Les lievres sauteurs sont strictement nocturnes. L’ensemble de la population presente une activitk maximale toutes les 24 heures. La taille du groupe et la distance au terrier sont en relation directe avec le temps CcoulC apres le coucher du soleil. Des tempera- tures basses ou moderkes ainsi que les grosses precipitations rtduisent I’activitC des lievres sauteurs, alors que le clair de lune limite la distance que I’animal parcourt sur les pans. Introduction The family Pedetidae consists of one species, the springhare, Pedetes capensis (Forster, 1778). The springhare is a large (3 kg), saltatorial-bipedal rodent which *Present address: Kibale Forest Project, P.O. Box 409. Fort Portal, Uganda. 7
Transcript
AJi. J . Ecol. 1984, Volume 22, pages 1-22
Nocturnal ecology of the springhare, Pedetes capensis, in Botswana
T. M. BUTYNSKI* Department of Wildlife, National Parks and Tourism, Gaborone, Botswana
Summary The springhare, Pedetes capensis, was studied in the Kalahari Desert, Republic of Botswana, from August 197 1 to August 1974 inclusive. Night censuses totalling 253 h provided data on springhare habitat selection, social grouping, activity pattern and the effect of moonlight and weather on activity.
Utilization of dry lakebeds (pans) by springhares was 2.3-3.4 times greater than on the surrounding habitats. Springhares were strictly nocturnal. The population, as a whole, exhibited one activity peak every 24 h. Group size and distance from burrows were directly related to time after sunset. Low temperatures and moderate to heavy precipitation both reduced springhare activity, while moonlight reduced the distance to which springhares moved out onto pans.
Except for adult males, all sex/age classes of springhares participated in group formation to the same extent. Apparently little social cohesion existed within groups. Territoriality was not observed but may occur in the area immediately around the burrow.
Re‘sume‘ Le lievre sauteur (Pedetes capensis) fut etudit dans le desert du Kalahari, au Botswana, de aoGt 1971 a aoGt 1974. Un total de 253 heures de recensements nocturnes a fourni des donnees sur le choix de I’habitat chez le lievre sauteur, son organisation sociale, ses types d’activites et l’effet du clair de lune et du temps sur celles-ci. L’utilisation du fond des lacs esskchtis (pans) par ces animaux est de 2.3 a 3.4 fois plus important que les habitats avoisinants. Les lievres sauteurs sont strictement nocturnes. L’ensemble de la population presente une activitk maximale toutes les 24 heures. La taille du groupe et la distance au terrier sont en relation directe avec le temps CcoulC apres le coucher du soleil. Des tempera- tures basses ou moderkes ainsi que les grosses precipitations rtduisent I’activitC des lievres sauteurs, alors que le clair de lune limite la distance que I’animal parcourt sur les pans.
Introduction The family Pedetidae consists of one species, the springhare, Pedetes capensis (Forster, 1778). The springhare is a large (3 kg), saltatorial-bipedal rodent which
is distributed over much of eastern and southern Africa (Dorst & Dandelot, 1970). This species is strictly nocturnal and spends the daylight hours within a burrow (Butynski & Mattingly, 1979). Springhares are monotocous, breed year-round and, although living in a strongly seasonal environment, exhibit no seasonality in breeding (Butynski, 1979; Butynski & Hanks, 1979) or moulting (Butynski, 1982) activity. Despite the wide distribution abundance and considerable econ- omic value of the springhare (Butynski, 1973) there is little information available on its nocturnal ecology.
Most studies of small, nocturnal mammals are based upon indirect data obtained from periodic trapping and laboratory experiments (Kaufman & Kaufman, 1982). Although such studies often contribute considerably towards an understanding of a species' behaviour, it is usually not known to what extent they reflect behaviour under truly natural conditions. Springhares, due to their large body size, high population densities, and preference for open habitats, offer an unusual opportunity to make direct observations on several aspects of the ecology and behaviour of a free-living nocturnal rodent.
This paper presents data on the above-ground activity pattern and habitat preferences of the springhare. Specifically, it examines springhare habitat selec- tion, social grouping, activity pattern and the effects of moonlight and weather on activity. The data presented here were collected from August 197 1 through August 1974 in the Republic of Botswana.
Study area Most of the observations were made in the Kutse Game Reserve and in the adjacent southern part of the Central Kalahari Game Reserve (23"-24"s and 24-25"E) (Dawson & Butynski, 1975). These areas are representative of the southern and central Kalahari bush savanna vegetation type (Smithers, 1971; van Rensburg, 1971; Wear, 1971) (Fig. 1).
The rainy season in the Kalahari is from October to April. Rainfall on the study area can vary 70% from the yearly mean of 400 mm. There is no perennial surface water in the region. Mean monthly maxima and minima air temperatures range from near 37°C and 11°C in January and December, to 27°C and -3°C in June and July.
The topography of the study area is predominantly flat or gently undulating. Loose aeolian sands of low fertility cover approximately 99% of the study area.
Four habitat types are recognized within the study area: (1 ) pan, (2) pan thicket, (3) bushveld or savanna, and (4) sand dune (Fig. 2).
Widely scattered throughout the region are circular, flat-bottomed, seasonally flooded depressions called 'pans'. Pans on the study area lie 5 m or more below the surrounding plain and range from 0.1 to 1.0 km in diameter. Pans are characterized by compact, clayey, calcareous soils and by bordering sand dunes. The dense, short grass cover, and paucity of woody vegetation on the pans strongly contrast with the scattered bush and tall grass on the surrounding bushveld. Due to differential utilization of the pan surface by springhares, three zones were arbitrarily chosen for detailed study at distance of 5 , 100 and 250 m from the edge of the pan (Table 1).
All vegetation not found on the pan surface is located on loose sandy soils.
Nocturnal ecology of the springhare 9
Botswana
A Zambia
Zimbabwe
N
Namibia
Francis t ow n. Central
Pans
Kutse Game *LePhePe - Central and southern Kalahari
bush savannah
.Tshane
/
Ghanzi Serowe
Kgalagadi
Kanye Park
"'"'";y Republic a1 South Africa
0 100 200 - w Kilometers
Fig. 1 . Map of the Republic of Botswana showing the location of the Kutse and Central Kalahari Game Reserves and other places mentioned in this paper.
A well-defined band ( 1 0-1 OOm wide) of dense woody vegetation surrounds most pans. This band is referred to as the 'pan thicket'. Plant species composition and structure within the pan thicket is distinct from both that found on the surface of pans and on the bushveld.
Most of the vegetation on the study area is classified as bushveld. Vegetation here is a mosaic created by differences in the relative abundance ofa small number of grass and woody species. There are few springhares on the bushveld, and thus, for the purposes of the present study, it is most useful to divide this extensive habitat into zones on the basis of distance from the nearest pan thicket, i.e. 10 m, 0.25 km, 1 km, and over 2 km (Fig. 2).
Sand-dune vegetation, although similar to that of the lower lying bushveld, tends to exhibit the taller and coarser grass species and the taller tree species.
10 T. M. Butynski
Pan Bushveld
Fig. 2. Schematic drawing of nine vegetation zones on the Kalahari study area and their relative use by springhares as determined by counts of fecal pellets (n=98 transects) and night censuses (n=489 springhares). (See text for details.)
Methods Line point transects (Riney, 1963) were used to obtain data on relative springhare use, soil type and composition of the vegetation within each vegetation zone (Table 2). Rhiney (1963), Caughley (1964) and Child (1968) all found this tech- nique suitable for assessing and comparing semi-arid vegetation types and their relative utilization by herbivores. Each transect consists of fifty points and fifty circular milacre (123-cm-radius) plots spaced two spaces (c. 3 m) apart. A notch located on the toe of the operator’s left boot is used to locate each point. The presence or absence of vegetation is assessed both above and below his notch. The placement of the notch also defines the centre of the plot within which all springhare fecal pellects are counted. A total of ninety-eight transects were run during the dry season and ninety-six during the wet season.
Springhares were counted during the night. A spotlamp connected to a 12-volt battery was manipulated by an observer located on the back of a truck. Censuses were made on pans and along the roads connecting the pans. While on the pans, the truck maintained a distance of approximately 50 m from the pans’ edge and a speed of 20 kph. This ensured that all portions of the pan were adequately searched. Springhare eyes glow particularly brightly when caught by light. In flat, open country, such as on pans, springhares were readily located at distances of 300 m or more. The ‘dazzling effect’ of the light usually made it possible to approach to within 20 m of the springhares. Springhares generally remained at the place where they were first encountered until the necessary data were col- lected. These data consisted of the distance from the edge of the pan at which each springhare was first sighted, group size, and the time and location of the
Nocturnal ecology of the springhare 1 1
Table 1. Predominant plant species associated with each ofnine vegetation zones in the Kalahari study area
Vegetation Woody Grass zone($ species species
250, 100 and 5 m from pan edge
Pan thicket
10 m, 0.25 km, I km, and 2 km or more from the nearest pan thicket
Baekea africana Aristida meridionalis Ochna pulchra Aristida uniplumis
Gre wia Jla va Schmidt ia pappop horoides Lonchocarpus nelsii Eragrostis lehmanniana
Terminalia sericea Ziziphus mucronata
sighting. The distance of springhares from the edge of the pans was estimated visually by three or four experienced observers.
Grass height on pans varied depending on season, rainfall and grazing inten- sity. This however, probably had no effect on the numbers of springhares counted since springhares alerted by the noise of the census vehicle moved to an upright position. In this position their eyes were nearly 40 cm above the surface of the pan and readily noticed by the observers. On only one occasion was a springhare seen to lower itself to the ground and apparently attempt to hide rather than remain in an upright posture or hop towards its burrow.
Night censuses on the bushveld differed from those on the pan in that the maximum distance at which springhares could be detected was reduced. Even so, it is probable that few of the springhares within 100 m of the observers went uncounted. Width of the pan transects was limited by the diameter of the pan while the width of bushveld transects was determined by the effects of the taller vegetation on visibility. In both cases the mean width ofthe transect approximated 200 m.
12 T. M. Butynski
Table 2. Wet- and dry-season vegetation composition and springhare use of Kalahari habitats
No. of Average springhare
Vegetation tree pellets zone and No. of Woody Grass Basal Grass Woody height milacre season transects litter litter grass canopy canopy (m) plot-'
250 m on to Pan
Dry season 100 m on to pan
Dry season Wet season
5 m on to Pan
Dry season Wet season
Pan thicket
Dry season Wet season
10 m into bushveld
Dry season Wet season
0.25 km into bushveld
Dry season Wet season
1 km into bushveld
Dry season Wet season
2 km into bushveld
Dry season Wet season
Dry season Wet season
Sand dune
6
I I 16
12 14
12 15
10 1 1
I I 4
9 8
14 25
13 3
0.0
0.0 0. I
0.5 0.0
19.8 14.5
5 .O 8.5
7.3 3 .O
10.2 4.4
10.4 8. I
19.7 9.3
6.3
14.5 16.6
7.7 16.4
11.5 15.7
13.0 21.6
18.5 9.0
12.4 7.8
7.2 8.2
18.6 12.0
16.0
17.8 20.4
10.8 22.7
2.8 4.4
4.0 5.3
7.8 8.0
6.4 8.8
5.3 8.6
4.3 6.7
33.3
55.3 72.7
18.2 58.8
34.7 39.2
74.8 42.9
52.4 39.5
40.7 44.2
23.2 40.2
43.1 44.7
0.0
0.9 0.0
0.2 0. I
3 7.0 39.2
19.2 15.6
12.9 9.5
14.0 11.2
15.3 10.8
17.5 13.3
0.0
0.0 0.0
0.0 0.0
1.8 1.8
I .3 2.1
I .5 I .o
I .6 I .o
1 . 1 1.4
I .7 1.3
0.23
2.14 -
4.72 -
0.39 -
I .69 -
0.64 -
0.16 -
0.88 -
0.46 -
Night censuses on the Kalahari study area were made monthly from January 1972 through August 1973, except for May 1972. The mean number ofcensuses per month was 2.25 f0-62 (SD). The mean number of h census-' was 3.3 f 1.40. An additional eighteen censuses were carried out on the Kalahari study area from September 1973 through August 1974. A total of twenty-five censuses were conducted in the Kgatleng district of eastern Botswana and in the Nxai Pan and Chobe National Parks of northern Botswana. Overall, eighty-five censuses, averaging 3.0 f 1.30 h census-', were conducted.
In most cases, springhare social groups were distinct and it was easy to deter- mine which animals were associated with one another and which were not. Only
Nocturnal ecology of the springhare 13
in areas of high springhare densities did it occasionally become difficult to deter- mine if a springhare was associated with a group. An arbitrary distance of 30 m between springhares was chosen to indicate whether two springhares were of the same group. Other workers have also found it necessary to use an arbitrary distance in their studies on group size (Caughley, 1964; Frith, 1964)
Cloud and wind conditions were recorded during censuses. Minimum tem- perature for the Kalahari study area was interpolated each night from data available for the two nearest Botswana Weather Bureau stations: Lephepe, 140 km to the east, and Letlhakeng, 110 km to the south, and also from some readings made at the Kutse Game Reserve. The South African Astronomical Observatory provided charts and graphs from which the time of sunset, moonrise, and moonlight intensity were calculated.
Springhares were located at night with the aid of a spotlamp and shot with a 12-gauge shotgun. An attempt was made to collect all springhares encountered. A total of 3 19 springhares were collected on the Kalahari study area and 24 1 on the Kgatleng study area. The mean monthly sample size from September 197 1 through July 1973 for both study areas combined was 24.2f 1.6 springhares.
Results Springhare utilization of pans The density of springhares on pans was considerably greater than in any other habitat. During night censuses the mean number of springhares seen per hour was twenty-one on the pans and nine off of the pans (Mann-Whitney U test, n, =32, n, = 4 I , U = 1 145, P < 0.000 1 ). Six pellet-count transects were randomly selected from the set of transects run in each of the nine vegetation zones. The mean number of pellets plot-, along the eighteen transects run on pans (2.36 pellets) was significantly higher than the mean number of pellets plot-' along the thirty-six transects conducted off the pans (0.70 pellets) (M-W, n, = 18, n,=36, U = 180, P < 0.005).
The edge of the pan was used more by springhares than any other part of the pan surface. Springhares sighted on pans during night censuses were placed into one of seven 'distance categories' according to their distance out from the edge of the pan when first sighted. Only censuses conducted during periods without any moonlight were used in this analysis. The number of springhares (x) sighted in each distance category was transformed by log, (x) in order to normalize the values. Likewise, the mean number of fecal pellets plot-' Q found along each of the twenty-nine transects run at 5 , 100 and 250 m on to pans was transformed by log, ( y + l ) . Night censuses (r=-0.90, df=5, P<O.OI) and pellet counts (r= -0.62, df=27, P<O.OOl) both indicate that springhare utilization of the pan surface decreased significantly with distance out from the edge of the pan (Fig. 2). Pellet counts suggest that springhares spent more than twice as much time along the edge of the pan than at 100 m from the edge.
Numbers of springhares counted on pans and the distance out on to pans to which springhares moved did not exhibit any pronounced monthly or seasonal patterns.
14 T. M. Butynski
Hours after sunset
Fig. 3. Relationship between time aRer sunset and mean number of springhares seen during night censuses on Kalahari pans.
Daily activity and feeding pattern Springhares were considered to be active when they were above ground and to be inactive when in their burrows. The time intervals between springhare sightings during night censuses provide an index to the daily activity pattern of springhares.
Springhares were strictly nocturnal and were not seen above ground during the daylight or crepuscular hours. Emergence from the burrow did not occur prior to 30 min after sunset and burrows were re-entered no later than 30 min before sunrise.
The springhare population, as a whole, exhibited one nightly peak of activity. The mean number of springhares sighted h-1 during night censuses was calculated for each hourly period throughout the night. The number of springhares on pans increased steadily during the first 5 h after sunset, remained more or less constant during the middle 4 h of the night, and decreased during the 4 h prior to sunrise (Fig. 3). This pattern of activity persisted throughout the year.
Springhares moved further out onto pans as the night progressed. The mean distance of springhares out from the edge of the pans was calculated for every 15-min period after sunset (n = 1045 springhares). Only data collected during periods without any moonlight were used in this analysis. Mean distance of springhares out from the pan edge increased significantly as the night progressed (r=0.62, df=25, P<O.OOl) . Mean distance of springhares from the pan edge increased from 32 m at 1 h after sunset to 96 m at 10 h after sunset.
Springhares left their burrows at the start of the activity period with empty or near-empty stomachs. Oven-dried weights of springhare stomach contents
Nocturnal ecology of the springhare 15
50
300 351
;/
*J - Y=9,3X -25.0
O'! Ib Ib 1; ;2 2'6 ;O ;4 Quarter hours af ter sunset
Fig. 4. Relationship between mean dry weight of springhare stomach contents and time after sunset ( n = 550 springhares).
showed significant positive correlation with the time (in 15-min periods) after sunset at which springhares were collected (r=0-96, df=23, P<0.0005) (Fig. 4).
Effects of moonlight on activity Activity of springhares on pans was influenced considerably by moonlight. Intensity of moonlight was assessed using a chart provided by the South African Astronomical Observatory. The chart relates intensity of moonlight to moon phase. Intensity of moonlight ranged from from 0% (no moon) to 100% (full moon). Percentage moonlight (x) during censuses was transformed by the function a. The distance of springhares out on to pans was inversely correlated with intensity of moonlight ( r = -0-95, df=6, P<O.OOl) (Fig. 5) . Under full-moon conditions the mean distance of springhares from the edge of the pan was about 4 m and few individuals were found more than 20 m from the edge. Mean distance of springhares out from the edge of the pan was 58 m when no moonlight was present.
Springhares fed further out on to pans when clouds reduced the amount of moonlight reaching the surface of the pan. During those twenty-six censuses which included both periods of moon presence and periods of moon absence (i.e. the moon rose and/or set during the census), springhares were a mean 38 m on to the pans when under moonlight as opposed to a mean 5 8 m in the absence of moonlight. This difference is highly significant (M-W, n, =20, n,=26 , U=381, P < 0.004).
16 T. M. Butynski
6or .
d(%) moonlight + 0 . 5
Fig. 5. Relationship between mean distance of springhares on to Kalahari pans and intensity of moonlight (n = 1480 springhares).
Table 3. Occurrence of springhare groups by size on Kalahari pans on moonlit and moonless nights ~~~ ~ ~~ ~
There was no significant difference between the proportion of springhares in social groups of two or more animals when moonlight was present (34%) and when it was absent (38%) (x*= 1.375, df= 1 , P>0-20) (Table 3).
As indicated above there was a strong negative relationship between the intensity of moonlight and the distances to which springhares moved out on to the pans. It it, therefore, somewhat surprising to find no significant correlation between intensity of moonlight and the number of springhares ( n =98 1 ) seen on pans during night censuses (r=0.14, df=30, P>O.lO).
Eflects of temperature and precipitation on activity As air temperatures decreased so did springhare activity. The number of spring- hares seen h-1 of census on pans ( n = 9 8 l ) was significantly correlated with the
Nocturnal ecology of the springhare I7
minimum air temperature during the nights on which censuses were conducted (r=0.41, df=27, P<0.025). At temperatures below 1°C few springhares were active.
Springhares remained active under conditions of light rainfall but moved into their burrows when precipitation became moderate to heavy.
Home range Those springhares associated with pans were frequently seen 25-250 m, and occasionally as far as 400 m, from their burrows. The shape and size of the home ranges were determined by the position of the burrow relative to the feeding area, and the size and quality of the feeding area. Thus, the home ranges of those springhares which fed on Kalahari pans were larger and more ‘T-shaped’ than the more circular home ranges of individuals which had burrows located right on the feeding area, as was often the case in eastern Botswana. The occurrence of springhare feeding groups indicates considerable home-range overlap, especially in dense populations.
Social groups Springhares actively formed groups. Thirty-six percent of the springhares on pans and 35% of springhares off pans occurred in groups (Table 4). A significant difference existed between the proportion of springhares that foraged in groups
Table 4. Occurrence of springhare groups by size in four areas of Botswana
Group size ~ ~~
I 2 3 4 5 6 Total
Kalahari pans No. of groups O/o of all groups No. of springhare O/o of all springhare
Kalahari bushveld No. of groups O/o of all groups No. of springhare O/o of all springhare
Kgatleng District No. of groups Yo of all groups No. of springhare O/o of all springhare
Northern Botswana No. of groups O/o of all groups No. of springhare O/o of all springhare
No. of groups Yo of all groups No. of springhare O/o of all springhare
Fig 6. Size of springhare groups plotted on a normal probability scale against their cumulative frequencies of occurrence.
in the Kalahari (36%) and in the Kgatleng District of eastern Botswana (47%)
Groups of springhares ranged in size from two to six animals. There was no significant difference between the relative frequency of groups of each size on the pans and in the bushveld (x2= 1.98, df=4, P>0.25). Data on the frequency of occurrence of different-sized groups of springhares for several parts of Botswana (n=2545 springhares) were combined and transformed by log, (y+ 1). There was a strong inverse relationship between the size of the group and the frequency of occurrence of each group size ( r = -0-93, df=4, P<0.005).
There is a weak, but significant, positive correlation between time after sunset (in 15-min periods) and mean size of springhare groups (r=0.34, df=32, P < 0.03).
No significant association was found between mean size of springhare groups and their distances out on to pans (r=0.26, df=5, P>0.20).
The cumulative frequency of occurrence of different-sized springhare groups was plotted on a normal probability scale against numbers per group (Cassie, 1954; Caughley, 1964) (Fig. 6). The points fell on a straight line with the exception of a deflection between points representing singletons and groups consisting of two springhares.
(xz= 17.78 df= I , P<0*005).
Composition of social groups In this analysis the data from springhares collected in the Kalahari and in the Kgatleng District were combined in order to increase the size of the sample. Male and female springhares were equally likely to occur in groups (x2=0-002, df= 1 , P>0.80). In a comparison among the five agehex categories (adult males,
Nocturnal ecology of the springhare 19
immature males, non-pregnant adult females, pregnant adult females, immature females), it was only adult males which were represented in groups more than expected by chance (x2=4.07, df= 1, P<0.05).
For groups consisting of but two individuals, the number of sexually homogen- ous groups did not differ significantly from the number of sexually heterogenous groups (x*=0.75 with Yate’s correction, df= 1, P>0.30), nor was the number of all male groups significantly different from the number of all female groups (x2=0.125 with Yate’s correction, df= 1, P>0.70).
Discussion Springhare utilization of pans Habitats with the highest utilization should represent near optimum conditions for the species (Dice, 1931; Jewell, 1935). The most suitable springhare habitats are flat open areas supporting a short grass cover with little or no woody veg- etation, and having sandy soils for burrowing. In northern and eastern Botswana the floodplains of rivers and swamps best meet these criteria. In the Kalahari, pans provide the best habitat.
There are three apparent reasons why springhares utilize pans relatively more than the bushveld. First, during all seasons, springhare food plants on the pans have a higher protein, mineral and water content than do those on the bushveld (Butynski, unpubl. data). Second, the flat surface of the pan and its short grass cover are probably optimal conditions for predator detection and avoidance by springhares. Third, these conditions may facilitate social behaviours such as finding mates.
Daily activity and feeding pattern From direct observations of free-living springhares, from the steadily increasing weight of stomach contents throughout the night, and from the increasing dis- tances at which springhares are found out on to the pans as the night progresses, it can be inferred that springhares have but one activity peak every 24 h. This activity pattern is similar to that described for several other rodents (Ashby, 1972). There is no evidence to suggest discontinuous feeding or the presence of more than one activity peak in the springhare, as is the case in many rodents (Miller, 1955; Brown, 1956; Reynolds, 1960; Kikkawa, 1964; Bergstedi, 1965; Cross, 1970; Blaustein & Fugle, 1981).
Efects of moonlight and temperature on activity Although activity of most mammals is more or less fixed from day to day, it is not inflexible. Certain weather conditions and moonlight are known to be particularly disruptive to activity of small mammals and must be taken into consideration when interpreting activity and population data.
Intensity of moonlight is of considerable importance as a factor affecting the activity of many nocturnal mammals. In the case of rodents, moonlight may inhibit activity (Blair, 195 1; O’Farrell, 1974; Lockard & Owings, 1974a; Kaufman
20 T. M. Butynski
& Kaufman, 1982), stimulate activity (Pearson, 1960; Owings & Lockard, 197 l ) , or have no apparent effect (Orr, 1959; Chew & Butterworth, 1964; Jorgensen & Hayward, 1965). Although moonlight inhibits springhare movements on pans, it does not reduce the amount of time which springhares spend on pans.
As the moon rises in the sky, and as the shadows produced by trees and shrubs become reduced in size, springhares move closer to trees and shrubs in order to remain within shadows. Similar behaviour has been recorded for Peromyscus leucopus (Falls, 1968), Dipodomys merriami (Justice, 1960), and Dipodomys ordii (Kaufman & Kaufman, 1982). It is likely that this ‘shadow-seeking’ behaviour by springhares is an adaptation for predator avoidance.
Springhare activity is suppressed under temperatures approaching freezing and almost totally curtailed at temperatures below freezing. The most obvious adaptive function of this behaviour on springhares and other rodents (Blaustein & Fugle, 1981; Vickery & Bider, 1981) is that it conserves energy and helps avoid thermoregulatory stress.
Size and composition of social groups Group size in free-living small mammals has been poorly documented. Although it seems that groups may offer members an advantage in predator detection and avoidance (Treisman, 1975; Bertram, 1978), there is no concrete evidence for this in small mammals. Group formation in small mammals, including spring- hares, may be most significant as a means by which reproductive, learning and foraging activities are facilitated.
Cloudsley-Thompson’s ( I 96 I ) observation that ‘truly’ social animals are seldom nocturnal, but nocturnal species may form aggregations for feeding ...’ applies well to springhares. It may be that, although springhares could benefit from a more complex social organization, it is not possible, under nocturnal conditions, to maintain the degree of contact between individuals that this would require.
There appears to be little social cohesion within springhare groups. In most cases, individuals join and leave groups without any apparent reaction from other members. Indications are that springhares prefer to feed in groups but tend to do so only when it is ‘convenient’ as, for example, around those food supplies and burrowing sites where animals are concentrated.
In Fig. 6 there is a deflection in the graphed line where it represents single animals and groups of two springhares. This deflection suggests that single animals tend to avoid being alone. The near absence of deviation from a straight line over the rest of the graph probably indicates that the mean group size is directly related to, and dependent upon, springhare densities.
The proportionate number of immature animals in springhare groups suggests that young springhares initiate above-ground activity with little attendance by adults. The low incidence of milk in the stomachs of immature springhares also shows that they rapidly become independent of adults upon leaving the burrow (Butynski, 1979).
Direct observations on behaviour and group formation indicate that, except in low-density situations, springhares have widely overlapping home ranges. They also suggest that actively defended territories do not exist, or at least are not associated with an area outside the immediate vicinity of the burrow.
Nocturnal ecology of the springhare 2 1
Acknowledgments Deepest appreciation goes to Dr G. Petrides for his guidance and encourage- ment, and to Drs R. Baker, J. King, D. Ullrey, Ms Jan Kalina, Ms M. Slavin and Mr R Kasul for their critical review of this paper. Mr A. Campbell, Director, Botswana Department of Wildlife and National Parks, Dr W. von Richter, F A 0 Wildlife Ecologist and Mr L. Birch, Chief Game Warden, deserve special thanks for their support. My warmest thanks are extended to Ms C. Wong, Ms A. Longenecker, Mr D. Longenecker, Mr. J. Dawson, Mr G. Mann, Mr D. Massey, and Mr ‘Kutse’ Aaron for their keen interest and assistance through- out this study. Lastly, I wish to thank Mr A. Anderson, Director of Botswana Meteorological Services, and Mr J. Swanepoel, Secretary of the South African Astronomical Observatory, for providing meteorological data. This paper is the result of a cooperative effort between the Botswana Department of Wildlife and National Parks, the United States Peace Corps and Michigan State University.
References ASHBY, K.R. (1972) Patterns ofdaily activity in mammals. Mammal Review, 1, 52-62. BERCSTEDI, B. (1965) Distribution, reproduction, growth and dynamics of the rodent species Clethrio-
nornys glareolus (Schreber), Apodemus flavicollic Welchoir) and Apodemus sylvaticus (Linnie) in southern Sweden. Oikos, 16, 132-1 50.
BERTRAM, B.C.R. (1978) Living in groups: predators and prey. In: Behavioural Ecology: an Evolution- ary Approach (Eds J. R. Krebs and N. B. Davies), pp. 64-96. Blackwell Scientific Publications, Oxford.
BLAIR, W.F. (195 I ) Population structure, social behaviour and environmental relations in a natural population of the beach mouse (Peromyscus polionotus leucocephalus). Contrib. Lab. Vert. Biol.. Univ. of Michigan, 48, 1-47.
BLAUSTEIN, A.R. & FUGLE, G.N. (1981) Activity patterns of Reithrodontomys megaloiis in Santa Barbara, California. J. Mammal. 62, 195-199.
BROWN, L.E. (1956) Field experiments in the activity of the small mammals, Apodemus. Clethrio- nomys and Microtus. J. Mammal. 126, 549-564.
BUTYNSKI, T.M. ( 1 973) Life history and economic value of the springhare (Pedetes capensis Forster) in Botswana. Botswana Noies and Records, 5,209-2 13.
BUTYNSKI, T.M. (1979) Reproductive ecology of the springhares Pedetes capensis in Botswana. J. Zool., Lond. 189,221-232.
BUTYNSKI, T.M. (1982) Pelage and molting in the springhare Pedetes capensis in Botswana. A,/;. J.
BUTYNSKI, T.M. & HANKS. J. (1979) Reproductive activity in the male springhare Pedetes capensis
BUTYNSKI, T.M. & MATTINCLY, R. (1979) Burrow structure and fossorial ecology o f the springhare
CAUCHLEY, G. (1964) Social organization and daily activity ofthe red kangaroo and the grey kangaroo.
CASSIE, R.M. (1954) The uses of probability paper in the analysis ofsize frequency distributions. Aust.
CHEW, R.M. & BUTTERWORTH. B.B. (1964) Ecology of rodents in Indian Cave (Mojave Desert), Joshua
CHILD, G. (1968) An Ecological Survey of Northeastern Botswana: A Report to the Government of
CLOUDSLEY-THOMPSON, J.L. (I96 I ) Animal Behaviour. Macmillan, New York. CROSS, R.M. (1970) Activity rhythms of the harvest mouse Micromvs minutus (Pallas). Mammalia,
DAWSON, J. & BUTYNSKI, T.M. (1975) The Kutse Game Reserve: preserving the Kalahari ecosystem,
ECd. 20, 279-287.
in Botswana. S. Afr. J. Wildl. Res. 9, 13-17.
Pedetes capensis in Botswana. A j . J. Ecol. 17,205-2 15.
J. Mammal. 45, 429436.
J. Marine Freshwater Res. 5 , 53 1-532.
Tree National Monument, California. J. Mammal. 45, 203-225.
DICE, L.R. (1931) Methods of indicating the abundance of mammals. J. Mammal. 12, 376-381. DORST, J. & DANDELOT, P.D. (1970) A Field Guide to the Larger Mammals ofAfiica. Collins, London. FALLS, J.B. (1968) Activity. In: Biology of Peromyscus (Ed. J. A. King), pp. 543-570. Special
FRITH, H.J. (1964) Mobility ofthe red kangaroo, Megaleia rufa. C.S.I.R.O. Wildl. Res. 9, 1-19. JEWELL. M.E. (1935) An ecological study of the freshwater sponges of northeastern Wisconsin. Ecol.
JORGENSEN, C.D. & HAYWARD, C.L. (1965) Mammals ofthe Nevada test site. Brigham Young Univ.
JUSTICE, K.E. (1960) Nocturnalism in Three Species of Desert Rodents. Ph.D. thesis, University of
KAUFMAN, D.W. & KAUFMAN, G.A. (1982) Effect of moonlight on activity and microhabitat use by
KIKKAWA, J. (1964) Movement, activity and distribution of the small rodents Clelhrionomy g/areo/us
LOCKHARD, R.B. & OWINGS, D.H. (1974a) Seasonal variation in moonlight avoidance by bannertail
LOCKARD, R.B. & OWINGS, D.H. (1974b) Moon related surface activity of bannertail (Dipodomys
MILLER, R.S. (1955) Activity rhythms in the woodmouse, Apodemus sylvaticus and the bank vole,
O'FARRELL, M.J. (1974) Seasonal activity patterns of rodents in a sagebrush community. J. Mammal.
ORR, H.D. (1959) Activity of white footed mice in relation to environment. J . Mammal. 40,2 13-22 I . OWINGS, D.H. & LOCKARD, R.B. (1971) Different nocturnal activity patterns of Peromyscus
PEARSON, O.P. (1960) Habits of harvest mice revealed by automatic photographic recorders. J.
VAN RENSBURG, H.J. (I97 I ) Range Ecology in Botswana. U.N.D.P. Tech. Document No. 2. Gaborone,
REYNOLDS, H.G. (1960) Life history notes on Merriam's kangaroo rat in southern Arizona. J.
RINEY, T. (1963) A rapid field technique and its application in describing conservation status and
SMITHERS, R.H.N. (1971) The Mammals of Botswana. Museum Memoir, 4. Mardon Printers,
TREISMAN, M. (1975) Predation and the evolution of gregariousness. I. Models for concealment and
VICKERY, W.L. & BIDER, J.R. (1981) The influence of weather on rodent activity. J . Mammal. 62,
WEAR, P.R. (1971) Vegetation in the Kalahari in Botswana. Botswana Notes and Records (pp. 88-95).
Publication of the American Society of Mammals, No. 2.
Monogr. 5,461-504.
Sci. Bull., Biol. Ser. 6 , 1-81.
Arizona.
Ord's kangaroo rat (Dipodomys ordii). J. Mammal. 63, 309-3 12.
and Apodemus sylvaticus in woodland. J . Anim. Ecol. 33, 259-299.
