ORIGINALARTICLE

Twentieth century demographic changesin cirio and cardon in Baja California,Mexico

Stephen H. Bullock1*, Nora E. Martijena1, Robert H. Webb2 and Raymond

M. Turner2

1Departamento de Biologıa de la Conservacion,

Centro de Investigacion Cientıfica y de

Educacion Superior de Ensenada, Baja

California, Mexico and 2US Geological Survey,

Tucson, AZ, USA

*Correspondence: Stephen H. Bullock,

Departamento de Biologıa de la Conservacion,

Centro de Investigacion Cientıfica y de

Educacion Superior de Ensenada, Apartado

Postal 2732, 22800 Ensenada, Baja California,

Mexico. E-mail: sbullock@cicese.mx

ABSTRACT

Aim Our purpose was to discern long-term large-scale patterns of survivorship

and recruitment of two dominant, charismatic and useful desert plants, cirio

[Fouquieriaceae; Fouquieria columnaris (Kell.) Kell. ex Curran] and cardon

[Cactaceae; Pachycereus pringlei (S.Watson) Britton & Rose], and to evaluate the

effects of physical and human factors.

Location The study included 77 sites distributed over c. 13,200 km2 in the

northern Vizcaıno Region of the Sonoran Desert, state of Baja California, Mexico.

Method Censuses used identified individuals (n ¼ 3780 cirio and 2246 cardon)

and were based on repeat photography. Time intervals between censuses ranged

from 29 to 96 years, ending in 1996–2002. Earlier repetition (1963) of old

photographs at 14 sites allowed analysis of temporal variation. The effect of

elapsed time was modelled with Weibull functions for survivorship and linear

functions for recruitment. To explain the residual variations, we tested categorical

and continuous variables related to substrate, geomorphology, climate and

geography. Human impacts were weighed with historical and quantitative

analysis.

Results Elapsed time accounted for 45% of the variation of survivorship in cirio

but only 35% in cardon. The fitted Weibull curves were used to estimate

longevity: all-size cohorts of 1000 individuals would die out in 388 (cirio) and

723 years (cardon). Recruitment in cirio was linearly related to time (r2 ¼ 0.71)

and averaged < 0.003 year)1 but was less than mortality at 90% of the sites. For

cardon, recruitment averaged > 0.005 year)1, was linearly related to time

(r2 ¼ 0.66) and was less than mortality at only 33% of the sites. Vital rates

were not strongly correlated between mid- and late-century but were mostly

similar to the long-term regional rates. Removing the factor of elapsed time, the

residual variations showed that survivorship was greater on sites protected from

winds for cirio and was less on very coarse soils for cardon. Recruitment increased

with latitude in cirio and was greater on southern exposures for both species. The

residual variations were not clearly affected, at a regional scale, by other factors

thought to be important, including elevation, distance to the Pacific Ocean,

geology, slope gradient, soil stability, older vs. young alluvial soils, and soil

Ca : Mg and Na : K ratios. Human impacts have been sporadic and

heterogeneous but locally strong; our quantitative indices of accessibility did

not show regionally significant effects. Blowdown by hurricanes is a sporadic

cause of major mortality for cirio but not necessarily for cardon.

Main conclusions At our scales, effects of time outweigh those of substrate, and

human impact was scant. Large patterns were pronounced: cirio experienced a

prolonged, widespread decline in the 1900s, while cardon fluctuated in different

directions and degree among local populations. Cirio was found to be inherently

Journal of Biogeography (J. Biogeogr.) (2005) 32, 127–143

ª 2005 Blackwell Publishing Ltd www.blackwellpublishing.com/jbi 127

much shorter-lived than cardon. We also suggest that recruitment in cirio was

low due to long periods with relatively dry winters that affected the entire region,

while spatial heterogeneity of cardon recruitment seemed more related to the

variation in summer rains.

Keywords

Climatic variability, demography, Fouquieria, human impact, hurricanes, land-

scape, Metapopulation, Pachycereus, Sonoran Desert, substrate.

RESUMEN

Metas Nuestro proposito fue discernir, en escalas de largo plazo y

geograficamente amplias, patrones de sobrevivencia y reclutamiento de dos

plantas carismaticas, utiles y dominantes del desierto: el cirio [Fouquieriaceae;

Fouquieria columnaris (Kell.) Kell. ex Curran] y el cardon [Cactaceae; Pachycereus

pringlei (S.Watson) Britton & Rose]. A su vez, evaluar los efectos de factores

fısicos y humanos sobre estos patrones.

Localizacion El estudio incluyo 77 sitios distribuidos en c. 13,200 km2 en el

norte de la Region Vizcaıno del Desierto Sonorense, en Baja California, Mexico.

Metodo Con censos basados en la repeticion de fotografıas se identificaron los

individuos (n ¼ 3780 cirios y 2246 cardones). Los intervalos de tiempo entre los

censos variaron entre 29 y 96 anos, terminando en 1996–2002. La repeticion en

1963 de viejas fotografıas en 14 sitios, permitio el analisis de la variacion

temporal. El efecto del tiempo transcurrido fue modelado con funciones

Weibull para la sobrevivencia y con funciones lineares en el caso del

reclutamiento. Para explicar las variaciones residuales, se pusieron a prueba

variables categoricas y continuas relacionadas a substrato, geomorfologıa, clima y

geografıa. Los impactos humanos se consideraron mediante analisis historico y

cuantitativo.

Resultados El tiempo transcurrido explico el 45% de la variacion en la

sobrevivencia del cirio pero solamente el 35% en la del cardon. El ajuste a la

funcion Weibull permitio estimar la longevidad: las cohortes de todos los

tamanos de 1000 individuos morirıan en 388 (para el cirio) y 723 anos (para el

cardon). El reclutamiento del cirio estuvo linearmente relacionado al tiempo

(r2 ¼ 0.71) y promedio < 0.003 ano)1 pero fue menor que la mortalidad en el

90% de los sitios. Para el cardon, el reclutamiento promedio fue de > 0.005

ano)1, estuvo linearmente relacionado al tiempo (r2 ¼ 0.66) y fue menor que la

mortalidad en solamente el 33% de los sitios. Las tasas vitales no estuvieron

fuertemente correlacionadas entre mediados y finales del siglo pero fueron

principalmente similares a las tasas regionales de largo plazo. Removiendo el

factor del tiempo transcurrido, las variaciones residuales mostraron que la

sobrevivencia fue mayor en sitios protegidos del viento para el cirio y fue menor

en suelos muy gruesos para el cardon. El reclutamiento incremento con la latitud

para el cirio y fue mayor en las exposiciones sur para ambas especies. Las

variaciones residuales no fueron claramente afectadas, a una escala regional, por

otros factores considerados como importantes, incluyendo la elevacion, la

distancia al Oceano Pacıfico, la geologıa, el gradiente de pendiente, la estabilidad

del suelo, la caracterizacion de los suelos aluviales en jovenes vs. viejos o las

proporciones de Ca:Mg o Na:K en el suelo. Los impactos humanos han sido

esporadicos y heterogeneos pero localmente muy fuertes; nuestros ındices

cuantitativos de accesibilidad no mostraron efectos regionalmente significativos.

S. H. Bullock et al.

128 Journal of Biogeography 32, 127–143, ª 2005 Blackwell Publishing Ltd

INTRODUCTION

In arid lands, moisture in the upper soil is only adequate for

the growth of most plants at sporadic intervals and then only

for short periods. Changes in individuals are often difficult to

perceive in these conditions, as are the trends of change in

populations. Some individual-based population studies have

extended to periods of several years to decades, but it remains

difficult to define both the spatio-temporal patterns of change

and their causes (Watson et al., 1997; Turner et al., 2003).

Large, succulent-stemmed species have been particularly well-

studied in Arizona, where populations of some of these and

other common or dominant species are far from static, at least

at the local level (Shreve, 1910; Niering et al., 1963; Tschirley &

Wagle, 1964; Steenbergh & Lowe, 1983; Goldberg & Turner,

1986; Parker, 1987,1988; Turner, 1990; Bowers et al., 1995;

Bowers, 1997; Pierson & Turner, 1998; Turner et al., 2003).

The dynamics of these populations and their ecosystems have

been affected strongly by human activity, directly and

indirectly (Niering et al., 1963; Bahre & Hutchinson, 1985;

Humphrey, 1987; Bahre & Shelton, 1993) with few exceptions

(Turner, 1990; Bowers et al., 1995). Climate variation also has

been a major factor (Brum, 1973; Turner, 1990; Pierson &

Turner, 1998; Turner et al., 2003), and there is evidence or

suggestions of other biotic or endogenous factors (Tschirley &

Wagle, 1964; Steenbergh & Lowe, 1983; McAuliffe, 1988).

Contrasts in population dynamics across substrates have

also been exemplified (Niering et al., 1963; Parker, 1987;

McAuliffe, 1994; Bowers, 1997; El-Keblawy et al., 1997;

Pierson & Turner, 1998). For a locality in the Vizcaıno region,

McAuliffe (1991) showed that the size structures differed

between adjacent alluvial terraces with different soil character-

istics, for populations of three dominant perennials: cirio

[Fouquieria columnaris (Kell.) Kell. ex Curran, Fouquieria-

ceae], cardon [Pachycereus pringlei (S. Watson) Britton & Rose,

Cactaceae] and gobernadora [Larrea tridentata (Moc. & Ses.)

Cav., Zygophyllaceae]. However, the interpretations may be

confounded by the problems of age estimation (Bullock et al.,

2004) and the probable instability of the populations on scales

of decades or centuries (see below). Substrate age can be

important, but a regional application could be confounded by

the diversity of geology, geomorphology and history.

In the context of the entire Sonoran Desert, or the world’s

deserts in general, the Vizcaıno Region of the Baja California

peninsula (Shreve, 1964) has been much less affected by

humans due to the scant physical and social infrastructure and

to the rough terrain and scarcity of water. Nonetheless, human

impacts cannot be neglected, nor can the persistent demands

to exploit some large perennial plants; technical assessments

are still needed to guide consideration of management options

(INE-SEMARNAP, 1997).

Cirio and cardon, woody succulents which often exceed 8 m

height (Fig. 1a), are two of the dominant and charismatic

plants of the Vizcaıno region. Both species have been subjected

to undocumented extraction for useful and decorative wood

and as ornamental plants. Both species are obliging subjects for

long-term measurements of growth, recruitment and mortality

due to their columnar form, but biological and population

data have been insufficient to evaluate their status or to

establish use or conservation plans. Current estimates of

longevity for cirio are 200–400 years (Bullock et al., 2004),

which is considerably less than previous estimates (Humphrey

& Humphrey, 1990; McAuliffe, 1991), while cardon reproduc-

tion may not start until age 100 years (M. Salazar & S.H.

Bullock, unpubl. data).

The present report explores long-term survivorship and

recruitment trends for these species. The study is unique for its

resolution, based on identified individuals, in combination

with a broad geographical and time scope, including 77 sites

scattered over c. 2� of latitude (Fig. 2), with census intervals of

29–96 years. It was possible to achieve this combination

through the use of repeat photography (Rogers et al., 1984;

Fig. 1a,b). We then evaluate several factors, which may affect

differences among sites in population dynamics, including

El derribamiento por huracanes fue una causa esporadica de mayor mortalidad en

el cirio pero no necesariamente en el cardon.

Principales conclusiones En nuestras escalas, los efectos del tiempo pesan mas

que los del substrato y el impacto humano es escaso. Los grandes patrones son

pronunciados: el cirio experimento una prolongada y amplia disminucion en los

1900s, mientras el cardon fluctuo en diferentes direcciones y grados entre

poblaciones locales. Se encontro que el cirio es inherentemente una planta de

muchamas corta vida que el cardon. Tambien sugerimos que el reclutamiento en el

cirio fue bajo debido a largos periodos con inviernos relativamente secos que

afectaron la region entera, mientras que la heterogeneidad espacial en el

reclutamiento del cardon parecio corresponder a la variacion en la lluvias de verano.

Palabras claves

Metapoblaciones, paisaje, demografıa, variabilidad climatica, huracanes,

impacto humano, substrato, Desierto Sonorense, Fouquieria, Pachycereus.

Demographic changes in cirio and cardon

Journal of Biogeography 32, 127–143, ª 2005 Blackwell Publishing Ltd 129

time, the history of human impacts, geographical situation and

substrate conditions. We explicitly compared survivorship and

recruitment rates between the early vs. late twentieth century,

and between soils derived from granite vs. basalt and old vs.

new alluvium, because of their contrasting hydrology and

nutrient status (McAuliffe, 1991; Graham & Franco-Vizcaıno,

1992; Franco-Vizcaıno et al., 1993; Franco-Vizcaıno, 1994).

We also consider the impacts of hurricanes and of long-term

variations in precipitation.

MATERIALS AND METHODS

Species

Cirio has the general form of a tapered column; its greatest

reported height is 26.4 m (Humphrey, 1991). The column is

sustained by lignified tissue forming a fenestrated tube. Water-

storage tissue within the tube may develop large lacunae as a

result of drought. Orthotropic branches are usually absent or

sparse. Plagiotropic branches are non-succulent and have

xeromorphic xylem, in contrast to the trunk (Carlquist, 2001).

These lateral branches have leaf-bearing short-shoots, are

abundant and may be several decimetres long on small plants

but are typically extremely reduced after 2–3 m above ground

(Henrickson, 1975). Leaves can expand at any time of year, with

adequate moisture (Humphrey, 1974); stem water reserves may

contribute to maintaining leaf turgor (Nilsen et al., 1990).

Photosynthesis is all C3; stem photosynthesis may contribute to

maintenance respiration and the initial growth of new leaves

(Franco-Vizcaıno et al., 1990). The flowers are hermaphroditic,

open in summer and are visited by a variety of insects (Nabhan

et al., 1999). The seeds are dispersed by wind in late fall and

establishment is probably limited to the first winter or spring

(Humphrey, 1974).

Cardon is a columnar cactus that can reach 20 m height

(Turner et al., 1995). The column and the lesser or greater

abundance of large, orthotropic branches, are sustained by a

set of woody xylem strands that are separate or modestly

interconnected (Niklas et al., 1999). Variations in water

storage causes the stem to shrink or swell radially. Respiration

is sustained by CAM photosynthesis in the stem surface

(Franco-Vizcaıno et al., 1990). The reproductive biology of

cardon is complex and geographically variable. Samples from

two sites in our area were gynodioecious (Fleming et al., 1998;

pers. obs.), and experiments in a trioecious population in

Sonora showed hermaphrodites were self-compatible (Flem-

ing et al., 1994). Flowering is in spring; pollinators include

bats, birds and insects (Moran, 1962; Fleming et al., 1994).

Seeds are dispersed in summer by fruit-eating animals;

establishment may be restricted to summer and fall (Cancino

et al., 1993).

Our study area included about half the latitudinal range of

cirio and certainly the larger part of its total population. The

latitudinal range of cardon is considerably greater, extending

sparsely 1.5� north of our area, and almost 6� further south

(based on Turner et al., 1995). Both species also occur in

coastal Sonora; cirio occurs in an extremely limited area and

cardon is quite broadly distributed (Turner et al., 1995).

Geography and climate

The study area is within the Mexican state of Baja California

between 30.123 and 28.696�N latitude (Fig. 2) and includes a

large part of the federal protection area for flora and fauna

‘Valle de los Cirios’ (30–28�N; Diario Oficial de la Federacion,

2 June 1980, Primera Seccion, pp. 9–10). The great majority

(91%) of our sites are on the Pacific slope of the peninsula and,

on the north–south axis, are distributed from the San Vicente

and San Fernando watersheds to the San Borja watershed (see

Fig. 2 inset). This segment of the peninsula is c. 90 (to 120) km

wide, with most of the land surface above 200 m but only small

areas above 1000 m on the eastern and southern limits. The

elevation of our sites ranged from 55 to 780 m, with 58.4%

between 400 and 700 m.

Figure 1 Matched photographs of a site near San Fernando

mission, Baja California, featuring the columnar succulents cirio

[Fouquieria columnaris (Kell.) Kell. ex Curran] and cardon

[Pachycereus pringlei (S.Watson) Britton & Rose]. (a) 1905

(E. A. Goldman, courtesy of the Smithsonian Institution);

(b) 1998 (S. H. Bullock, CICESE).

S. H. Bullock et al.

130 Journal of Biogeography 32, 127–143, ª 2005 Blackwell Publishing Ltd

The geological framework of the area consists of intrusive

felsic plutons (tonalites to granodiorites) and associated

metasedimentary rocks, extrusive basalts and intrusive ultra-

mafic rocks, marine sedimentary rocks, and Quaternary basin

fills (Gastil et al., 1975). Development of drainage networks is

chaotic, leading to a landscape patchwork of through-flowing

drainages and a few closed basins with playas. Relatively gentle

topography is of limited extent and diverse origin: a deposi-

tional plain SE of San Fernando; an erosional surface of the

batholith NW of Catavina; a rolling, basalt littered terrain NW

of Chapala; and a large valley with eroded marine sediments,

north and south of Punta Prieta.

Regional surveys of the soils include coarse scale mapping

with broad classification units and only scattered points of

verification (DETENAL, 1982). Fine-scale topographic vari-

ation hampers use of the soil maps for characterizing small

sites. However, the most widespread classes in our area are

lithic and rocky (INEGI, 2001). Soils are rarely sodic or

alkaline (DETENAL, 1982; Graham & Franco-Vizcaıno, 1992;

INEGI, 2001; E. Franco-Vizcaıno, pers. comm.). Caliche

(pedogenic carbonate) and thick petrocalic deposits are

prominent in some areas (INEGI, 2001) but have not been

well studied. Basalt-derived soils are notable for high clay

content, and granite-derived soils have high sand content. One

interesting feature in relation to the diversity and abundance of

succulent plants, is the variation in Ca:Mg ratio, although this

does not correspond to different parent rocks (Franco-

Vizcaıno et al., 1993), and so is not yet mappable.

The area is part of the Sonoran Desert as delimited according

to vegetation (Shreve, 1942) and climate (Schmidt, 1989) and

corresponds to the northern part of the Central Desert of the

Baja California peninsula as delimited by Aschmann (1959). In

terms of Shreve’s (1964) regions of the Sonoran Desert, the

study area corresponds to most of the northern Vizcaıno

region. Along with the Sonoran biogeographical affinity,

elements of a broad transition from/to California are common

(Shreve, 1936; Brown et al., 1992; Bullock, 1999).

Mean annual precipitation varies from115 to 153 mmamong

stations, but is non-normal in both seasonal and annual totals

(Bullock, 2003). Precipitation occurs predominately in the cool

season (61–91% in November–April) and for this season the

history of variation is highly correlated among stations in our

area (Bullock, 2003). Warm-season precipitation, of tropical or

low subtropical origin, can be extremely high but is usually very

local. The tropics also contribute occasional hurricanes (Court,

1980), which are perhaps the most erratic and unique feature of

this region’s climate; their high winds may be important to the

tall plants we studied. Available data showed no relation between

elevation and precipitation, but the contrast of Pacific and gulf

slope precipitation and the frequent incursion of coolmarine air

and even fog from the Pacific, produce an uneven gradient

inland from the west coast that strongly affects the peninsula’s

botanical characteristics (Shreve, 1964;Hastings&Turner, 1965;

Humphrey, 1974).

Reconstruction of cool-season precipitation from proxy

records indicated that exceptionally wet years were not only

Figure 2 Maps of the study area, with regional context, reference localities and census sites.

Demographic changes in cirio and cardon

Journal of Biogeography 32, 127–143, ª 2005 Blackwell Publishing Ltd 131

rare but very unevenly distributed over the last 150 years: only

20 years between 1850 and 1997 had high precipitation

(January through March anomalies > 2.0 at San Diego), and

nine of these occurred after 1975 (Bullock, 2003). The recent

concentration of wet cool seasons raises the larger question of

whether the distribution of precipitation variability is station-

ary at the scale of > 50 years or if there have been larger shifts

or directional changes (Webb & Betancourt, 1992; Giorgi

et al., 1994; Meko et al., 1995; Turner et al., 2003). Attempts to

reconstruct the history of warm-season precipitation have not

been successful. Atmospheric and oceanic processes underlying

climatic patterns for the region have been discussed by various

authors (e.g. Garcıa & Mosino, 1968; Markham, 1972; Webb &

Betancourt, 1992).

Palaeoecology

The history of the biota of the Vizcaıno region is still difficult

to sketch. As recently as 10,000 years bp, pinyon–juniper

woodland occupied sites at Catavina and San Fernando

(Penalba & Van Devender, 1998; Wells, 2000; Sankey et al.,

2001) and a juniper–chaparral vegetation occupied a site south

of Sta. Gertrudis (Rhode, 2002); these sites are now dominated

by desert species. Pleistocene megafaunal fossils have been

found near Chapala (L. Davis, pers. com.) as well as further

south on the peninsula (Ferrusquıa-Villafranca & Torres-

Roldan, 1980); this fauna is unknown in the peninsula’s rock

art, some of which may predate 5000 years bp (Fullola et al.,

1993), although human occupation is evident from before

9000 years bp (Davis, 2003). In addition, repeated strong

changes of sea surface temperatures during at least the last

6000 years, which are evidenced by faunal changes in human

middens of marine molluscs (Moriarty, 1968), might have

affected the incursion of tropical storms, the oceanic moder-

ation of peninsular temperatures, and the occurrence of fog.

During this long period of human occupation, major climatic

fluctuations can be inferred from variations in lake levels of

now-dry playas (Arnold, 1957; Davis, 2003).

Human population and land-use

The pre-evangelization Cochimı population between the

territory of the San Ignacio mission (established 1728) and

San Fernando was estimated at 11,100 (Aschmann, 1959),

representing a density of c. 0.22 persons km)1. The economy

was entirely based on hunting and gathering of terrestrial and

marine resources. Concentration of the Cochimı at the

missions was impressive: San Borja had a maximum popula-

tion of c. 1800, and San Fernando reached c. 1400 (Aschmann,

1959, p. 182). However, the indigenous population declined

rapidly due to disease; by 1836, 15 years after Mexican

independence, the region’s mission population was less than

100 (Aschmann, 1959).

Secular occupation of the peninsula by people of mestizo or

European culture began at the end of the Jesuit period (1768),

unlike the early occupation (with livestock) of the eastern and

north-eastern areas of the Sonoran Desert (Turner et al.,

2003). Colonization was extremely scant North of Sta.

Gertrudis until the short-lived mining booms of the late

1800s, particularly the gold rush in the Calmallı district

starting in 1882 (Gabb, 1867; Deasy & Gerhard, 1944; Chaput

et al., 1992).

Many of the old ranches of the region are associated with

families that Martınez (1965) reported as immigrating to the

peninsula in the 1700s, but a historical account of the ranches

has yet to be made. Certainly the population was extremely

small until colonization efforts were promoted and subsidized

by the Mexican government in the 1970s and 1980s; the

network of dirt roads was vastly extended, and the density

(if not the dispersion) of ranches increased. In the period

1967–1979, the federal government established collective

landholdings (ejidos) that occupied almost the entire study

area (Fig. 3). The transpeninsular highway was first opened as

a dirt road in 1927, and was somewhat rerouted, graded and

paved in 1970–1973. A recent review estimated the population

of our area as roughly 572 individuals, with 50 permanent

settlements and 15 sporadically occupied localities, excluding

coastal fishing settlements that are more populous (Aramburo

et al., 2002). However, between the crash of the missions and

1930, censuses record only one settlement with > 100 inhab-

itants, the village at the onyx mine ‘El Marmol’ (c. 1930; Deasy

& Gerhard, 1944; Espinoza Arroyo, 1992). As of the year 2002,

Figure 3 Twentieth century historical events, including the

temporal distribution of old photographs (no. of sites recorded),

winter rainfall anomaly at San Diego [(observed ) mean)/SD],

territorial expansion of collective land holdings (ejidos), years with

tropical storms and hurricanes affecting the study area, years of

construction of the unpaved and paved transpeninsular highways,

years of ‘Baja 1000’ off-pavement vehicle races, and years of

exploitation of palmilla (Yucca schidigera Roezl ex Ortgies).

S. H. Bullock et al.

132 Journal of Biogeography 32, 127–143, ª 2005 Blackwell Publishing Ltd

there were three settlements of > 100 inhabitants in our area

(Catavina, Punta Prieta and Rosarito).

Livestock were introduced more than 230 years ago (del

Barco, 1988, p. 304), but have remained extremely localized on

this arid landscape. After the establishment of the ejidos

(1967–79; Fig. 3), attempts at cattle ranching increased,

sustained by liberal subsidies from the federal government;

these in turn floated on the high price of petroleum. The rural

subsidy programmes did not fare well through the national

economic crises of the 1980s and 1990s. Goats have been

exceptionally rare in our area, although abundant and

devastating in the mountains south of Sta. Gertrudis. Feral

burros probably have been present but are not conspicuous

and their numbers have never been estimated. The lack of

surface water has limited the cattle population and restricted

most of its activity to moister arroyos and their immediate

vicinity, where impacts can be intense, but no careful study has

been made. Government range scientists classified the entire

region, except some major arroyos and the sierras SE of San

Borja, as unsuitable for any attempts at livestock ranching

(Paulın-Ramırez et al., 1981), but there have been many small,

marginal operations. In our area, the estimated number of

cattle in 1998 was < 7500 (Aramburo et al., 2002).

Census sites and methods

The sites for this study were chosen in an area relatively well

represented by old data, and that is relatively homogeneous in

climate and present and historical land use. Some potential

sites were excluded due to recent heavy disturbance, absence of

both cardon and cirio, poor quality of the old data or

uncertainties in the photointerpretation. In addition, an effort

was made to obtain a broad representation in terms of

geography and time.

The initial materials for this study were 107 old photographs

taken by 21 travellers and obtained from 15 archives,

individuals or publications (see Acknowledgments). The old

photographs dated from 1905 to 1967 such that the intervals

between censuses ranged from 29 to 95 years and were

> 55 years at half the sites (Fig. 3). At 14 sites, old photo-

graphs were repeated by J. R. Hastings and R. M. Turner in

1963, so that our work represented the third recording of those

sites (Fig. 2). In these cases, the earlier intervals were

25–40 years and the later interval was uniformly 33 years.

(Old documentary information is too vague for demography,

although occasionally useful for vegetation studies; Minnich &

Franco-Vizcaıno, 1998.)

The views were located in the field between 1996 and 2002.

Following standard methods (Rogers et al., 1984), the old

photographs were repeated from a camera position indistin-

guishable in the field from the old position. The old and new

photographs were printed to match. Censuses were then made

by detailed inspection, identifying and tabulating survivors,

disappearances (deaths) and appearances (recruits). In a few

cases where the number of individuals was small, the data were

combined from closely adjacent localities with similar substrate

and equivalent years of record. In addition, some photographs

were internally partitioned into two sites in order to separate

areas of highly contrasting substrates. The resulting 77 sites are

mapped in Fig. 2. We used the variation among sites to test for

factors affecting survivorship and recruitment, instead of

combining results among sites (Bowers et al., 1995).

We characterized several physical features of each site,

including latitude, elevation, distance from the Pacific coast

and precipitation. In order to approximate the variation of

precipitation, we grouped the sites into 10 physiographical

subareas, which could be characterized from the few weather

stations [nine reported by Bullock, 2003, plus Krutzio, on the

Pacific coast SSW of Punta Prieta (Adrian Aguirre Munoz,

pers. com.)]. Sites in eight subareas were assigned the mean

annual rainfall (1977+) of the included station. The ninth

subarea was assigned the mean of two adjacent stations, and

the 10th was given the mean of three stations.

We also developed several qualitative classifications of other

physical factors that might affect population dynamics:

landform (slope, pediment, plain, valley); geology (granite,

basalt, sedimentary, metasedimentary, mixed); topographic

protection from high winds (summit, flat, hills, valley); soil

texture (four classes); soil stability (eight classes); slope

gradient (four classes) and exposure (open, southerly, nor-

therly). Following the observations of McAuliffe (1991), we

compared vital rates between sites with old and new alluvial

soils. In addition, based on the observations of Franco-

Vizcaıno et al. (1993), we compared vital rates among the few

sites where detailed soil data were available.

Accessibility to people was described for each site in terms of

distances to road, ranch and town.

Demographic calculations

Survivorship was modelled by nonlinear regression with the

Weibull function

Nt ¼ N0 eð�t=BÞc

where N0 is 1000, t is time in years, and B and C are the fitted

constants. The observed number of survivors, of all sizes, at

each site was adjusted to an initial population of 1000 for this

analysis (Ebert, 1999); for either species, we excluded sites with

less than five individuals in the oldest photograph. The effects

of variables other than time were evaluated using the residuals

of survivorship (observed ) predicted from the Weibull

curves). This treatment differs substantially from Bowers et al.

(1995), who did not model a trend in survivorship but

calculated only cumulative mortality.

Recruitment was calculated as the ratio of new individuals to

the median between the initial number and the number of

survivors. Recruitment was modelled as a linear function of

time, and the residuals were used to test the effects of other

variables.

Residuals of the survivorship and recruitment regressions

were analysed with single factor analysis of variance or

regression, except the quantitative variables were also examined

Demographic changes in cirio and cardon

Journal of Biogeography 32, 127–143, ª 2005 Blackwell Publishing Ltd 133

together in linear models. The qualitative variables presented

problems of unequal numbers of categories and non-uniform

distribution of cases among categories; parametric and nonpar-

ametric tests were applied, and for some analyses categories were

censored or combined. Sample sizes (number of sites) were

particularly small for the soil chemistry variables.

For these analyses, the sites with censuses from three

different years were represented by the data from the first and

last censuses only.

Census errors

Censuses of visible individuals are readily made for columnar

(and many other) species based on ground-based photogra-

phy. Nonetheless, it is important to consider the relative

accuracy of censuses from photogrammetric and standard

plots. Potential errors derive from problems of resolution due

to differences between camera lenses or lighting conditions at

the times of the original and repeat photographs. These

problems generally can be avoided by excluding doubtful parts

of the view in both photographs. Similarly, some parts of a

view may have to be excluded due to obscuring vegetation in

the old or new photographs. Other characteristics of repeat

photography are a variance in the intensity of sampling (in any

size of plot) because of geomorphic obstructions, a high

variance in ‘plot size’ among sites, and a strong bias against

inclusion of the smallest plants (particularly those < 30 cm

tall). The recruitment and mortality we report is of well-

established plants, probably more than 5 or 10 years old (pers.

obs.). Age is indeterminate, however. At least in the case of

cirio, studies of growth rates have shown that size is a very

poor indicator of age (Escoto-Rodrıguez & Bullock, 2002;

Ramırez Apud Lopez, 2002; Bullock et al., 2004).

In the case of cirio, the occasional occurrence of basal

branching may lead to overestimation of abundance, but no

consistent bias regarding vital rates. The geographical and

temporal variation of multiple trunks has not been assessed,

but basal branching has not been detected at all sites, and is not

common where it has been observed. Apparently, growth rates

of individual trunks do not differ appreciably between solitary

and multiple-trunk plants (Escoto-Rodrıguez, 1999).

RESULTS

Human impacts

For prehispanic times in the Vizcaıno region, destructive use of

live plants parts other than seeds and fruits, was only remarked

for Agave (Aschmann, 1959). In the mission decades, collecting

wood for fuel and construction may have had locally

significant effects. Cirio was used heavily for firewood at San

Borja although the missionaries considered it a remarkably

useless plant (del Barco, 1988, p. 93) and the Cochimı

considered it harmful (Linck, in Burrus, 1967, p. 46; see also

del Barco, loc. cit.). Cardon has long been used for its edible

fruits.

In the short-lived mining era, large amounts of wood were

required for structural uses and for fuel in steam-powered

stamp mills and smelters (Chaput et al., 1992), particularly in

the Calmallı district (Southworth, 1889, plates on pp. 87–88),

at San Fernando (Bose & Wittich, 1912–1913, plate 86), at

Julio Cesar (Bose & Wittich, 1912–1913, plate 80) and at San

Juan. Cirio and cardon often may have been spared the axe due

to their succulence, but as at mission San Borja, demand was

extreme and resulted in some notable examples of very local

depletion (Aschmann, 1972). The limitation of mine produc-

tion by scarcity of wood was occasionally remarked (e.g.

Engerrand & Paredes, 1912–1913) and is underscored by the

frequent practice of shipping raw ore to Europe for processing

(Chaput et al., 1992). The populous and long-lasting settle-

ment at El Marmol fuelled its kitchens with locally abundant

mesquite, although some propane may have been used since

the 1950s (P. Maclish, pers. com.).

The post-1960s increase of human population in the desert

brought greater demands for wood for fuel and construction,

more cattle, and more commerce in local products with distant

markets. Wood from cardon and cirio have been widely used,

the former principally to support roofing (Moran, 1968), and

the latter for corrals (Fig. 4) and various ornamental uses,

including wall coverings, for which cirio has had a market

outside the region. Extensive land uses probably have not had

obvious effects on cirio or cardon. Cattle ranching has

persisted in marginal operations since the early 1800s and

was heavily subsidized in the 1970s and 1980s, but impacts on

the vegetation are obvious only in some areas with permanent,

quality forage or water. The gathering of fruits of jojoba

(Simmondsia chinensis (Link) Schneid.), which lasted for a few

years in the 1970s and 1980s (Bullock, 1999), entailed well-

dispersed but selective and seasonal incursions. In addition,

harvesting of stems of palmilla (Yucca schidigera Roezl

ex Ortgies) in the Vizcaıno region, which dates from the 1970s,

is limited to scattered dense stands of a few thousand hectares

to the NW of Catavina (Carranza-Acevedo, 1997; Bullock,

1999); these stands lack sizeable cirio or cardon populations.

Vegetation has been destroyed incidental to road-building

and urbanization. Off-road vehicle use, which is known

elsewhere to be extremely damaging to desert soils and

vegetation (Webb & Wilshire, 1983), has been unrestricted

(and unstudied). However, all-terrain races have been run on

existing roads except, perhaps, for informal and/or poorly

documented events in the 1950s and 1960s (Edith Bernal,

SECTURE, unpubl. data; Thawley, 1968; Cortes, 1995).

Hurricanes

Baja California is occasionally affected by hurricanes that have

generated in low latitudes of the Eastern Pacific, as in

September 1959 and September 1997 (‘Nora’). Thorough data

on Eastern Pacific storms, including tracking and wind

intensity, is relatively recent (Court, 1980) so the frequency

and geography of impacts are not well known. Historical data

show our study area was crossed by 15 tropical storms or

S. H. Bullock et al.

134 Journal of Biogeography 32, 127–143, ª 2005 Blackwell Publishing Ltd

hurricanes between 1920 and 2002 (Figs 3 & 5; Court, 1980;

Garcıa et al., 1980; http://weather.unisys.com/hurricane/

e_pacific). At least the 1959 and 1997 events traversed our

area at hurricane strength. Although the incursions between

1921 and 2002 give an average of 0.19 year)1, there were

periods of > 15 years without any identified incursion, but five

were reported in the 1920s (Garcıa et al., 1980). Of course,

hurricanes passing offshore or somewhat to the south of our

area also may have important impacts because the diameter of

the area of extreme winds may vary greatly (Hurd, 1929; Serra

Castelan, 1971).

Three surveys of Hurricane Nora’s impact were made in

November 1997. About 55 km ENE of Nora’s estimated

landfall (c. 93 km SSE of San Fernando) and 38 km north of

Punta Prieta, Clark & Ward (2000) observed 15% mortality in

cirio but none in cardon on a sandy 1.4 ha plot (n ¼ 31, 24,

respectively). In the same area, our own counts on more varied

soils along an 11.2 km transect showed c. 4.2% mortality in

cirio and 1.0% in cardon (considering plants > 1.5 m tall,

n ¼ 851, 796). We also made observations slightly further east

(c. 25 km NE of Punta Prieta) along a 12.6 km transect, and

found cirio was more affected than cardon (3.6% vs. 0.6%,

n ¼ 1452, 1470).

Photogrammetric plot data

We identified 3780 cirios and 2246 cardons at 77 sites, including

long-term survivors (2256 and 1508, respectively) and recruits

(404 and 491). Cirio occurred at more sites and was generally

more common than cardon. At the initial censuses (i.e. in the

older photographs), there were more than 50 individuals at

about one-third of the sites for cirio but at only c. 19%of the sites

for cardon. The maximum local sample sizes in the initial

censuses were 257 for cirio and 151 for cardon. There was a mild

positive relation between the abundances of the two species at

the chosen sites (Pearson’s r ¼ 0.339 in the old photos and 0.332

in the new, both P < 0.01; or, excluding four outliers, 0.500 and

0.514, P > 0.001).

Survivorship was significantly related to the time interval

between photographs. The fitted Weibull curves of survivors

(adjusted to an initial N ¼ 1000) vs. time interval explained

45% of the variation for cirio and 35% for cardon for sites with

N0 ‡ 5 (both P < 0.001, n ¼ 72 and 48 sites, respectively;

Fig. 6). Because the curves are nonlinear, no single, overall

value of survivorship can be given. However, as an indication

of longevity, we calculated the estimated time until an all-size

cohort of 1000 would become extinct: 388 years for cirio and

723 years for cardon.

For cirio, by restricting the analyses to sites with ‡ 15 or

‡ 50 individuals (n ¼ 62 and 26), the explicative power of

elapsed time alone increased to 49% and 63% of the variance,

respectively. However, the confidence limits for the parameters

were very similar at all three N0, and were narrowest at N0 ‡ 5.

For cardon, limiting the analysis to N0 ‡ 15 gave poorerFigure 5 Recorded tracks of tropical storms and hurricanes that

crossed part of the study area, 1920–2002.

Figure 4 A 1996 view of a corral constructed of c. 113 cirio trunks, adjacent to ruins of the San Fernando mission; previous photographs of

this view showed no corral (1923) and a corral made entirely of mesquite (Prosopis sp.; 1963).

Demographic changes in cirio and cardon

Journal of Biogeography 32, 127–143, ª 2005 Blackwell Publishing Ltd 135

confidence intervals but a similar r2 (34%, n ¼ 33 sites), and

the change in parameters showed a substantial change in the

survivorship pattern. At N0 ‡ 50 (n ¼ 13), the distribution of

points was inadequate for analysis. For both species, the curves

fitted with N0 ‡ 5 were used to calculate residuals for analysis

with other potentially contributing factors.

Survivorship residuals for cirio were significantly affected by

landform and protection, which had similar explanatory power

(r2 ¼ 0.125 and 0.116, F ¼ 3.20 and 2.97, P ¼ 0.029 and

0.038; anova). However, these factors were also highly

correlated (Pearson v2 ¼ 40.8, P < 0.001). Sites with excep-

tionally low survivorship were often on plains or without

topographic protection. For cardon, in contrast, the only

significant factor was soil texture. Due to the small number of

heavy clay sites with cardon, the analysis was restricted to the

other three conditions. This showed lower survivorship on the

coarsest soils [r2 ¼ 0.164, F ¼ 4.11, P ¼ 0.023 (the result with

all four textures was also significant, r2 ¼ 0.172, P ¼ 0.039);

anova]. For both species, the substrate factor was also tested

as contrasts limited to basalt vs. granite-derived soil and old vs.

new alluvium; there were no significant effects on survivorship.

Notably, the factors representing human access had no

regional significance in either species.

Recruitment rates from the regional regressions were

0.00282 (individuals/individual) year)1 in cirio and

0.00542 year)1 in cardon, with the exclusion of three outliers

in each species. Time accounted for 71% and 66% of the

variation among sites (P < 0.001, Fig. 7). The median and

maximum of time-corrected rates for individual sites were

0.00233 and 0.01010 year)1 for cirio, compared with 0.00517

and 0.08670 year)1 for cardon. Among the sites with cardon,

16% had recruitment rates exceeding 0.01 year)1. Recruitment

was less than mortality at 90% of the sites for cirio, compared

with only 33% of the sites for cardon.

Several factors contributed significantly to explaining the

residual variation of recruitment. In cirio, recruitment

increased with latitude (r2 ¼ 0.146, P < 0.001; linear regres-

sion). In addition, there was evidence of a relationship with

precipitation, although this was apparently nonlinear and

much weaker (r2 ¼ 0.045, P ¼ 0.045), and precipitation was

correlated with latitude (Pearson’s r ¼ 0.695, P < 0.001). In

both species, recruitment was greater on southern exposures

than open sites or northern exposures (cirio, r2 ¼ 0.107,

F ¼ 3.95, P ¼ 0.024; cardon, r2 ¼ 0.200, F ¼ 5.37, P ¼ 0.008;

anova). Among the substrate categories, there was one

significant contrast for cirio, with higher recruitment on

granitic than basalt-derived soils (r2 ¼ 0.128, F ¼ 4.56,

P ¼ 0.041; anova). However, this contrast was not significant

for cardon, and neither species showed differences between old

and new alluvial soils at the regional level. As for survivorship,

the factors representing human access had no regional

significance in either species.

Additional quantitative soil properties were available for six

sites (E. Franco-Vizcaıno, unpubl. data), which allowed tests of

effects on cirio (cardon was too rare at these sites). Neither

survivorship nor recruitment residuals were significantly

correlated with upper or lower soil Ca : Mg or Na : K ratios,

or with upper soil NO3 or PO4. The range in Ca : Mg ratios

among these sites was larger than that tested by Franco-

Vizcaıno et al. (1993) and the range of the Na : K ratio was

somewhat less.

The effects of the various factors that contributed signifi-

cantly to explaining the variation in survivorship and recruit-

ment among sites in either species are summarized in Table 1.

For a few sites, earlier repeat photographs (1963) permitted

the comparison of two periods of about three decades each.

Survivorship and recruitment rates for the two periods were

compared with each other and with the long-term data in

terms of their differences from the long-term trends for each

Figure 6 Observed and estimated (fitted Weibull function)

relationships between survivorship and time interval for cirio

(triangles and solid line) and cardon (circles and broken line).

Survivorship is expressed as number of survivors, adjusted to an

initial population size of 1000 (1000Nt /N0, where N0 is the

number in the first census (old photograph) and Nt is the number

of survivors and recruits in the new photograph).

Figure 7 Observed and estimated (fitted linear function) rela-

tionships between recruitment and time interval for cirio (triangles

and solid line) and cardon (circles and broken line). Recruitment

is Nr /((N0 + Ns)/2), where Nr is the number of recruits, N0 is the

number in the first census and Ns is the number of survivors.

S. H. Bullock et al.

136 Journal of Biogeography 32, 127–143, ª 2005 Blackwell Publishing Ltd

species (n ¼ 12 sites for cirio, n ¼ 6 for cardon). Survivorship

and recruitment ‘residuals’ for the two periods were not closely

correlated in either species. By nonparametric comparison

with the long-term residuals, cardon showed no significant

differences in the overall variation of survivorship or recruit-

ment. Cirio showed no significant differences in recruitment,

but survivorship in both periods was exceptionally low

compared with the long-term, all-site trend (Mann–Whitney

test, P ¼ 0.032 for first period vs. long-term, P ¼ 0.024 for

second vs. long-term).

DISCUSSION

During the 1900s, mortality exceeded recruitment for cirio at

90% of sites in the region of its greatest abundance. This

species, a charismatic indicator of the Sonoran Desert in Baja

California, experienced a decline that was large in numbers,

geographically extensive and long-term. For cardon, on the

contrary, recruitment was higher than mortality at 67% of its

sites, indicating that the regional population of this species has

been either stable or slightly increasing, in accord with studies

in Sonora (Turner et al., 2003). Geographically extensive

explanations for the trends are as yet speculative and are

limited to ideas about the impacts of humans and the physical

environment, particularly climate. There has been no attention

to interactions with other organisms such as nurse plants, pests

or herbivores.

These trends are balances and may find explanation on the

side of survivorship or recruitment or both. These processes

can be examined separately because our method used identi-

fied individuals, not simply abundances. As expected, both

these processes were strongly time-dependent, to the extent of

explaining 45–63% of the variation in survivorship among sites

in cirio and 35% in cardon, as well as 71% and 66% for

recruitment. The residual variations were explained in as much

as 20% by some physical factor, but the factors differed

between species and between processes. Only site exposure

affected both species in regard to recruitment.

Time alone remains the major variable, which suggests merit

in further study of survivorship and recruitment, particularly

including estimates of age or stage (Pierson & Turner, 1998).

However, such estimates are difficult at best and subject to

large errors at worst (Bullock et al., 2004). The all-size cohorts

used for our survivorship curves may include plants differing

in age by tens or hundreds of years. This is probably the reason

that elapsed time explained much more variation for

recruitment than for survivorship. At the same time, further

study is needed of mortality factors, the conditions of

establishment, and of variations in the balance between these

processes in both species. Further study by other methods is

clearly needed to improve the representation of small juvenile

plants.

The net result of mortality and recruitment is depicted in

relation to elapsed time in Fig. 8. These results suggest a

tendency for cardon to increase and cirio to decrease. Lacking

any particular conceptual model, the trends were adequately

treated as linear, with an increase of 0.116% year)1 in cardon

and a much stronger and faster trend in cirio, decreasing at

0.446% year)1.

Survivorship

At the scale of one century, survivorship was best modelled as

decreasing with time in cirio, but models with constant or

decreasing survivorship were very similar in cardon. The

strength of the models is remarkable considering the disper-

Table 1 Significant explanatory factors for

variation among sites of survivorship and

recruitment in cirio and cardon Factors

Cirio Cardon

Survivorship Recruitment Survivorship Recruitment

Elapsed time (Ni > 5) r2 ¼ 0.454*** 0.708*** 0.347*** 0.664***

Latitude 0.146***

Landform 0.125*

Exposure (S vs. N) 0.107* 0.200**

Substrate (texture) 0.164*

Substrate (basalt vs. granite) 0.128*

*P < 0.05, **P < 0.01, ***P < 0.001.

Figure 8 Net change in cirio and cardon (recruitment ) mor-

tality), expressed as the percentage of the initially observed pop-

ulation at each site. For cirio (triangles), net change ¼ ) 0.446.

(elapsed time) (r2 ¼ 0.680, F ¼ 151.2, P > 0.001); for cardon

(circles), net change ¼ 0.116. (elapsed time) (r2 ¼ 0.178, F ¼ 9.8,

P < 0.001).

Demographic changes in cirio and cardon

Journal of Biogeography 32, 127–143, ª 2005 Blackwell Publishing Ltd 137

sion and heterogeneity of the sites. Unfortunately, retrospec-

tive validation of the models cannot be extended significantly.

Our treatment of all sizes together, starting at an indefinite size

well beyond ‘seedling’, is a reasonable first approximation

because size apparently has a poor relation to age (Bullock

et al., 2004).

The models provide the first estimates of longevity in these

species and show a strong contrast between them. To estimate

local extinction time, the size of the initial cohort must be

specified; thus, for N0 ¼ 1000, extinction would be in less than

400 years for cirio but more than 700 years for cardon. The

times to halving of any N0 would be c. 71 and 140 years,

respectively. These estimates are not directly comparable with

other species for which long-term survivorship is not known

or has not been modelled numerically. Nonetheless, both cirio

and cardon are apparently much longer-lived than the large

saguaro cactus (Carnegiea gigantea (Englm.) Britton & Rose),

for which ages derived from growth rates suggest a ‘longevity’

of less than 200 years (Shreve, 1935; Turner et al., 2003), which

in turn is similar to the arborescent succulent kokerboom

(Aloe dichotoma Linn.f., Liliaceae) of south-western Africa

(Vogel, 1974), while falling short of the estimates of supra-

millenial longevity in gobernadora (McAuliffe, 1988).

Turner (1990) showed that mortality rates varied consider-

ably during the twentieth century for several large perennials in a

protected crater in the Pinacate region of Sonora, principally in

relation to drought. Those variations were important to long-

term demographic changes. In contrast, the saguaro population

at the same site did not show large variations in mortality, such

that its unstable age structure could be attributed to variations in

recruitment (Turner, 1990). However, lacking an analysis for

saguaro similar to that for cardon, we note again that time

accounted for only 35% of the variation in mortality of cardon,

so that the possibility of major dieback due to disease, for

example, remains of interest (Holguin et al., 1993).

Only two substrate variables contributed significantly to the

variation in survivorship, and these were not the same for cirio

and cardon. Cirio lost an exceptionally larger proportion of

individuals on sites offering no topographic protection from

high winds, while cardon mortality was higher on sites with

very coarse or very fine soils. At least in the context of our

design and data, several factors were not significant despite

suggestions by earlier authors, including the contrast of older

and newer alluvium, basalt and granite, some aspects of soil

chemistry, and slope exposition. McAuliffe (1991) found

contrasts in population structures of cirio and cardon between

alluvial surfaces of different ages at one site. The difference

with our negative result for surface age may be due to our

focus on observed population dynamics or problems with

McAuliffe’s assumptions about growth. However, our negative

result may be an artefact because our regional perspective

required that ‘new’ refer to very recent and ‘old’ include all

other surfaces, with a great variety of ages and histories. More

directly related to soil chemistry, our sparse data for effects

Ca : Mg, Na : K, NO3 and PO4, suggest no strong demogra-

phic effect that might correspond to the suggested effect on

succulent diversity (Franco-Vizcaıno et al., 1993). On the

contrary, the unexpected negative results for basalt vs. granite

corresponded to a negative result for soil texture in cirio. For

cardon, the parent rock contrast also was not significant, but

soil texture did explain some of the residual variation.

Like many columnar succulents, cirio and cardon can be

toppled by high winds. The impact of extreme north or east

winds has not been studied, but the passage of hurricanes over

the northern Vizcaıno is clearly a mortality factor of

importance (Clark & Ward, 2000; this study). The four to

six times greater impact of Hurricane Nora on cirio compared

with cardon may be largely a result of differences in the root

systems: cardon has both a near-surface, wide-spreading

system and a basal bayonet root (Niklas et al., 2002) but

cirio has only a shallow system with few major radials.

Although the long-term history of hurricane landfalls will

probably remain obscure, and many years may pass without a

landfall, the meteorology and inferred history of Eastern

Pacific hurricanes (Court, 1980) leaves no doubt of the

relevance of hurricanes to plant populations at the scale of

several decades and centuries.

Although cirio has been in demand for construction,

decoration and even fuel, observed mortality did not relate to

our indices of accessibility. Thus, we suggest that direct

exploitation has not had a simple or regionally significant

impact. Nonetheless, exploitation of the trunks is a locally

important factor even if the geography of impacts is complex.

Following the establishment of the ejidos in 1967–79, there

was undoubtedly increased use of cirio with the increases in

resident human population and cattle, although these remain

at low levels. No estimations of use have been attempted, but

cirio is still very common in homes, commercial establish-

ments and now-deteriorating corrals (Fig. 4). Moreover, just

as the ejidos flourished with the coming of the paved

highway, so did tourism and the consequent demand for cirio

wood for decoration, construction and export. Cardon ‘ribs’

are often used in construction, as small beams and posts in

roofs, walls and fencing, but apparently have no export

market.

With respect to the exploitation of wood of both species,

however, local people claim to use only naturally dead and

dried plants because unnatural death results in rotting; this has

not been independently verified. Because of its apparent

general decline, cirio populations should be protected from

excessive felling for local use or export.

Six of the sites were very near to missions, but none of these

showed an exceptional survivorship in terms of residuals of the

general time–survivorship relationship for cirio and cardon.

Photographs were also repeated at the San Fernando and Julio

Cesar copper mines, showing substantial increases in cover by

some shrubs, but the numbers of cirios and cardons distin-

guishable in the older photographs did not meet our minimum

criterion to include among the censuses. Recovery of cirio

around the San Fernando mine might require 200 years,

according to Aschmann’s (1972) interpretation of an informal

survey.

S. H. Bullock et al.

138 Journal of Biogeography 32, 127–143, ª 2005 Blackwell Publishing Ltd

Recruitment

Elapsed time explained c. 70% of the variation in recruitment

in both cirio and cardon. For several reasons, there is no

necessary implication that establishment was spatially or

temporally uniform. Of course, in our methods recruitment

is not seedling establishment but appearance, which integrates

establishment over longer periods. The strength of the

relationship was undoubtedly enhanced by the long intervals

between censuses and by the interval of 29 years shared by all

sites (i.e. 1967–1996). In addition, excluding a few extreme

outliers in each species greatly improved the relationships.

Surprisingly, the median annual rate of recruitment was more

than twice as high in cardon (0.5%, compared with 0.2% in

cirio), and the maximum rate was more than eight times

higher in cardon (8.7% vs. 1.0%).

Some of the residual variation in cirio was explained by

latitude. This may be partly an effect of precipitation, but the

analysis suggested that some other gradients must also be

involved. Despite this positive relationship with latitude, cirio

also showed better recruitment on southerly exposures, which

was also true for cardon. A similar trend for exposure has been

noted for large cacti in the northern part of their distribution

(Parker, 1987,1993; Pierson & Turner, 1998). Temperature

regimes in the Vizcaıno region have not been studied

sufficiently to determine if the exposure effect could be

attributed to killing frosts. Of course, field experiments on

establishment would be particularly useful if active manage-

ment becomes a priority. Factors other than climate that merit

attention are nurse plant populations (McAuliffe, 1988) and the

relation of individual growth rates to demographic processes.

Cirio establishment is supposed to be favoured by wet

winters followed by cool summers (Humphrey, 1974). In the

region of the sites reported here, the history of winter rainfall is

spatially more uniform than summer rainfall (Bullock, 2003).

Thus, establishment should tend to be regionally synchronized.

However, opportunities for establishment may have been very

limited in the last 100 years. In the period of climatic records

(< 50 years), only c. 1 year in 10 had a cool season with 3 or

even 2 months of > 50 mm precipitation. A proxy-based

reconstruction for nearly 150 years (Bullock, 2003) suggests

opportunities for establishment have been not only infrequent

but also very unevenly distributed in time, as nearly half of the

exceptionally wet winters occurred after 1975. On the basis of

size structure of one population, Aschmann (1972) suggested

there were two or three recruitment events from c. 1900 to

c. 1950.

The long-term variation in precipitation suggests that cirio

may have had increased establishment in the late 1900s. These

plants would not be apparent in the repeat photographs due to

their diminutive size and the disguising of cirio’s conical trunk

by an abundance of lateral branches in the ‘juvenile’ stage. The

existence of such cohort awaits an independent test.

In contrast to cirio, cardon establishment may be favoured

by wet conditions in summer and fall, resulting from

incursions of tropical air masses, perhaps followed by mild

winters. The evidence in favour of this idea is that seed

dispersal occurs in summer (Moran, 1968), seed germination is

lower at temperatures below 40 �C (Cancino et al., 1993), and

the species is more widely distributed in the southern portion

of the peninsula where precipitation occurs predominantly in

summer (Hastings & Turner, 1965; Turner et al., 1995).

Appreciable summer precipitation is uncommon at any place

in the northern Vizcaıno region, and most rainfalls are local, as

shown by the poor spatial–temporal coherence of summer

precipitation records (Bullock, 2003). The corresponding

pattern of establishment for cardon should be sporadic events

that are asynchronous among localities. This is apparent to

some extent in the outlying values of recruitment, reaching

more than 15 times the median, and perhaps in the

diversity among sites of population increases and decreases.

Nonetheless, it is remarkable that, overall, greater recruitment

and less mortality has occurred in the species which is more

responsive to summer rains.

In cardon, both the diversity among sites and the overall

majority of increasing populations is remarkably similar to

results from its coastal populations in Sonora (Turner et al.,

2003). Thus, cardon contrasts with saguaro, the other giant

columnar cactus of the Sonoran Desert, which appears to be in

a widespread, long-term decline (Turner et al., 2003). The

natural history of saguaro bears little resemblance to cirio, so

the causes of their declines probably differ. The potential

difficulty in understanding cirio and cardon is suggested by the

history and variety of hypotheses regarding changes in saguaro

(e.g. Niering et al., 1963), most recently including global

climate change (Thompson et al., 1998). Compared with

declines of arborescent succulents in south-western Africa

(Midgley et al., 1997), the decline of cirio is apparently on a

different time scale and is not attributable to obvious non-

climatic factors.

Populations of desert perennials may be typically unstable

(Turner, 1990; Van Devender, 1990; McAuliffe, 1994), and the

patterns and causes of their changes may be heterogeneous

among species at any site or among sites for any species

(Parker, 1987; Turner, 1990; McAuliffe, 1994; Nabhan &

Suzan, 1994; El-Keblawy et al., 1997; Turner et al., 2003). Cirio

and cardon show very different overall patterns at the same

sites across a substantial region. The decades-long decline of

cirio may be a consequence of regional pulses in establishment

during the past 400 years, with establishment in the twentieth

century failing to match the mortality of earlier cohorts. This is

analogous to interpretations of the fluctuations of other

perennials of various life forms, in more continental but

biogeographically related regions (Jordan & Nobel, 1981;

Goldberg & Turner, 1986; Parker, 1993; Turner, 1990; Bowers

et al., 1995; Pierson & Turner, 1998; Turner et al., 2003). As

with cirio and cardon, many of these species have shown net

changes that are similar although numerically disparate

between the early and late twentieth century (Bowers et al.,

1995; Turner et al., 2003). Episodic establishment and con-

tinuous mortality could provide a conceptual framework for

the population dynamics of cirio and cardon, but we cannot

Demographic changes in cirio and cardon

Journal of Biogeography 32, 127–143, ª 2005 Blackwell Publishing Ltd 139

yet rule out other combinations of continuous and episodic

processes (Watson et al., 1997). In addition, a regionally

coherent hypothesis like direct effects of climatic variability

ought to be subject to other tests (Jordan & Nobel, 1981;

El-Keblawy et al., 1997; Watson et al., 1997; Pierson & Turner,

1998), and need not apply at any particular site, considering

other idiosyncratic features.

Our methods permit contrasts of overall population trends,

geographical variation and factors affecting survivorship and

recruitment, based on known individuals at many sites over a

large region, observed at intervals of several decades to a

century. This study provides the first synchronic view of

vegetation processes in the coastal, subtropical desert of the

Vizcaıno region, showing time scales that contrast between two

dominant plant species and with shorter-lived dominants of

other regions. The apparent paucity of human impacts is

remarkable, but much variation remains to be explained in

terms other than our characterizations of the populations and

their physical environment.

ACKNOWLEDGEMENTS

We are grateful to the individuals whose photographs have

become a legacy of growing importance: C.G. Abbott, B.A.

Arnold, H. Aschmann, E. Bose, H. Bravo-Hollis, A. Bridge,

H.W. Crosby, N.H. Darton, J. Engerrand, E.A. Goldman, H.E.

Gulick, J.R. Hastings, L.M. Huey, G.E. Lindsay, T.D. Mallery,

M. McDonald, T. Paredes, E.S. Ross, F. Shreve, R.M. Turner,

and I.L. Wiggins. We also thank several individuals as well as

institutions and their personnel for facilitating the use of old

photographs: Brigham Arnold, Louise Aschmann, Helia Bravo-

Hollis, Arthur Bridge and Edward Ross; US Geological Survey

Desert Laboratory, Smithsonian Institution Archives, Special

Collections of the University of Arizona Libraries, Mandeville

Special Collections of the University of California, San Diego,

US Geological Survey Photographic Library (Denver), Library

of the San Diego Natural History Museum and the US

National Archives.

This project was funded by CICESE, the National Geo-

graphic Society (5576-95) and USGS, and was assisted by

sabbatical support from CONACYT (990293) and San Diego

State University. We are grateful for comments on the

manuscript by Paul Zedler, Paul Dayton, Kathleen Parker

and anonymous reviewers, and for technical support by Mario

Salazar, and by Alejandro Hinojosa, Cesar Almeda, Hugo

Riemann, Fernando Escoto and Sean Connin. Loans of

equipment and facilities were kindly made by USGS-Tucson,

Glenn Northcutt, Sergio Ramos, CICESE’s Department of

Optics, and L.G. Alvarez. Adrian Aguirre kindly provided

precipitation data from Krutzio. We thank the people of the

region for their tolerance and help.

REFERENCES

Aramburo, G., Espejel, I., Leyva, C. & Martınez, R. (2002)

Programa de manejo del area de proteccion de flora y fauna

valle de los Cirios, Baja California. Informe Tecnico Final.

Comision Nacional de Areas Protegidas, Secretarıa de Eco-

logıa Medio Ambiente y Recursos Naturales, Mexico.

Arnold, B.A. (1957) Late Pleistocene and recent changes in

land forms, climate and archaeology in Central Baja Cali-

fornia. University of California Publications in Geography, 10,

201–318.

Aschmann, H. (1959) The Central Desert of Baja California:

demography and ecology. Ibero-americana, 42, 1–315.

Aschmann, H. (1972) Recovery of desert vegetation. Interna-

tional geography, 1972, Vol. 1 (ed. by W.P. Adams and F.M.

Helleiner), pp. 631–633. University of Toronto Press, Tor-

onto.

Bahre, C.J. & Hutchinson, C.F. (1985) The impact of historical

fuelwood cutting on semidesert woodlands of southeastern

Arizona. Journal of Forest History, 29, 175–186.

Bahre, C.J. & Shelton, M.L. (1993) Historic vegetation change,

mesquite increases, and climate in southeastern Arizona.

Journal of Biogeography, 20, 489–504.

Bose, E. & Wittich, E. (1912–13) Informe relativo a la

exploracion de la region Norte de la costa occidental de la

Baja California. Parergones del Instituto Geologico de Mexico,

4, 307–533 (plates LXXXVIII–CXII).

Bowers, J.E. (1997) Demographic patterns of Ferocactus cyl-

indraceus in relation to substrate age and grazing history.

Plant Ecology, 133, 37–48.

Bowers, J.E., Webb, R.H. & Rondeau, R.J. (1995) Longevity,

recruitment and mortality of desert plants in Grand

Canyon, Arizona, USA. Journal of Vegetation Science, 6,

551–564.

Brown, J.W., Real, H.G. & Faulkner, D.K. (1992) Butterflies of

Baja California: faunal survey, natural history, conservation

biology. Lepidoptera Research Foundation, Beverly Hills, CA.

Brum, G.D. (1973) Ecology of the saguaro (Carnegiea gigan-

tea): phenology and establishment in marginal populations.

Madrono, 22, 195–204.

Bullock, S.H. (1999) La vegetacion del noroeste de Baja Cali-

fornia en el contexto de la inestabilidad ambiental. Revista

Chilena de Historia Natural, 72, 501–516.

Bullock, S.H. (2003) Seasonality, spatial coherence and history

of precipitation in a desert region of the Baja California

peninsula. Journal of Arid Environments, 53, 169–182.

Bullock, S.H., Turner, R.M., Hastings, J.R., Escoto-Rodrıguez,

M., Ramırez Apud Lopez, Z. & Rodrıguez-Navarro, J.L.

(2004) Variance of size-age curves: bootstrapping with

covariance. Ecology, 85, 2114–2117.

Burrus, E.J. (1967) Wenceslaus Linck’s reports and letters 1762–

1778. Dawson’s Book Shop, Los Angeles, CA.

Cancino, J., Leon de la Luz, J.L., Coria, R. & Romero, H.

(1993) Effect of heat treatment on germination of seeds of

cardon [Pachycereus pringlei (S. Wats.) Britt. & Rose, Cac-

taceae]. Journal of the Arizona-Nevada Academy of Science,

27, 47–54.

Carlquist, S. (2001) Wood anatomy of Fouquieriaceae in

relation to habit, ecology, and systematics; nature of meri-

stems in wood and bark. Aliso, 19, 137–163.

S. H. Bullock et al.

140 Journal of Biogeography 32, 127–143, ª 2005 Blackwell Publishing Ltd

Carranza-Acevedo, R.A. (1997) Caracterizacion ecologico-silvi-

cola de poblaciones de Yucca schidigera con potencial de

aprovechamiento. Master’s Thesis, Facultad de Ciencias,

Universidad Autonoma de Baja California, Ensenada.

Chaput, D., Mason, W.H. & Loperena, D.Z. (1992) Modest

fortunes: mining in northern Baja California. Natural History

Museum of Los Angeles County, Los Angeles.

Clark, W.H. & Ward, D.M. Jr (2000) Hurricane impacts in the

Central Desert of Baja California Norte, Mexico. Haseltonia,

7, 81–85.

Cortes, J.A. (1995) Bajo el desafio Baja 1000. Reforma,

November, 29, 1995, 3.

Court, A. (1980) Tropical cyclone effects on California. NOAA

Technical Memorandum NWS WR, 159, 1–45.

Davis, L.G. (2003) Geoarchaeology and geochronology of

pluvial Lake Chapala, Baja California, Mexico. Geoarchae-

ology, 18, 205–223.

Deasy, G.F. & Gerhard, P. (1944) Settlements in Baja Cali-

fornia: 1768–1930. Geographical Review, 34, 574–586.

del Barco, M. (1988) Historia natural y cronica de la antigua

California [adiciones y correciones a la noticia de Miguel

Venegas] (ed. by M. Leon-Portilla), 2nd edn. Universidad

Nacional Autonoma de Mexico, Mexico.

DETENAL (1982) Cartas Edafologicas 1 : 250000. Direccion

General de Geografıa del Territorio Nacional, Mexico.

Ebert, T.A. (1999) Plant and animal populations: methods in

demography. Academic Press, San Diego, CA.

El-Keblawy, A., Shaltout, K.H., Lovett-Doust, J. & Ramadan, A.

(1997) Population dynamics of an Egyptian desert shrub,

Thymelaeahirsuta.CanadianJournalofBotany,75,2027–2037.

Engerrand, J. & Paredes, T. (1912–13) Informe relativo a la

parte occidental de la region Norte de la Baja California.

Parergones del Instituto Geologico de Mexico, 4, 278–306

(plates LXVI–LXXXVII).

Escoto-Rodrıguez, M. (1999) Variaciones temporales y espaci-

ales en la tasa de crecimiento del cirio (Fouquieria col-

umnaris). Master’s Thesis, Facultad de Ciencias,

Universidad Autonoma de Baja California, Ensenada.

Escoto-Rodrıguez, M. & Bullock, S.H. (2002) Long-term

growth rates of cirio (Fouquieria columnaris), a giant suc-

culent of the Sonoran Desert in Baja California. Journal of

Arid Environments, 50, 593–611.

Espinoza Arroyo, A. (1992) Los Rosarenos: memorias del naci-

miento y vida de un pueblo bajacaliforniano, 1774–1992.

Sector Pesquero de El Rosario & Museo de Historia de

Ensenada, Ensenada.

Ferrusquıa-Villafranca, I. & Torres-Roldan, V. (1980) El

registro de mamıferos terrestres del Mesozoico y Cenozoico

de Baja California. Universidad Autonoma de Mexico, In-

stituto de Geologıa, Revista, 4, 56–62.

Fleming, T.H., Maurice, S., Buchmann, S.L. & Tuttle, M.D.

(1994) Reproductive biology and relative male and female

fitness in a trioecious cactus, Pachycereus pringlei (Cacta-

ceae). American Journal of Botany, 81, 854–867.

Fleming, T.H., Maurice, S. & Hamrick, J.L. (1998) Geographic

variation in the breeding system and the evolutionary sta-

bility of trioecy in Pachycereus pringlei (Cactaceae). Evolu-

tionary Ecology, 12, 279–289.

Franco-Vizcaıno, E. (1994) Water regime in soils and plants

along an aridity gradient in central Baja California, Mexico.

Journal of Arid Environments, 27, 309–323.

Franco-Vizcaıno, E., Goldstein, G. & Ting, I.P. (1990) Com-

parative gas exchange of leaves and bark in three stem

succulents of Baja California. American Journal of Botany,

77, 1272–1278.

Franco-Vizcaıno, E., Graham, R.C. & Alexander, E.B. (1993)

Plant species diversity and chemical properties of soils in the

central desert of Baja California, Mexico. Soil Science, 155,

406–416.

Fullola, J.M., Petit, M.A., Rubio, A., del Castillo, V. & Bergada,

M.M. (1993) Esquema crono-cultural del poblamiento

prehistorico de las sierras centrales de la penınsula de Baja

California, Mexico. Arqueologıa, 9/10, 3–15.

Gabb, W.M. (1867) Exploration of Lower California.

Resources of the Pacific Slope (ed. by J.R. Browne) [In

appendix A sketch of the settlement and exploration

of Lower California], pp. 82–112. D. Appleton & Co,

New York.

Garcıa, E. & Mosino, P. (1968) Los climas de la Baja Cali-

fornia. Comite Mexicano para el Decenio Hidrologico

Internacional. Memoria 1966–67, pp. 29–56. Instituto de

Geofısica, Universidad Nacional Autonoma de Mexico,

Mexico.

Garcıa, J.A., Castelan, S.S. & Betancourt, A.E. (1980) Trayec-

torias de ciclones tropicales. Centro de Ciencias de la

Atmosfera, Mexico. Universidad Nacional Autonoma de

Mexico, Mexico.

Gastil, R.G., Phillips, R.P. & Allison, E.C. (1975)

Reconnaissance geology of the state of Baja California.

Geological Society of America Memoir, 140, 1–170.

Giorgi, F., Shields Brodeur, C. & Bates, G.T. (1994) Regional

climate change scenarios over the United States produced

with a nested regional climate model. Journal of Climate, 7,

375–399.

Goldberg, D.E. & Turner, R.M. (1986) Vegetation change and

plant demography in permanent plots in the Sonoran

Desert. Ecology, 67, 695–712.

Graham, R.C. & Franco-Vizcaıno, E. (1992) Soils on igneous

and metavolcanic rocks in the Sonoran Desert of Baja

California. Geoderma, 54, 1–21.

Hastings, J.R. & Turner, R.M. (1965) Seasonal precipitation

regimes in Baja California, Mexico. Geografiska Annaler,

47A, 204–222.

Henrickson, J. (1975) Vegetative morphology of the cirio

(Fouquieria columnaris). Cactus and Succulent Journal,

(Suppl.), 33–46.

Holguin, G., Bowers, R. & Bashan, Y. (1993) The degeneration

of cardon populations in Baja California Sur, Mexico.

Cactus and Succulent Journal, 65, 64–67.

Humphrey, R.R. (1974) The boojum and its home: Idria col-

umnaris Kellogg and its ecological niche. University of Ari-

zona Press, Tucson, AZ.

Demographic changes in cirio and cardon

Journal of Biogeography 32, 127–143, ª 2005 Blackwell Publishing Ltd 141

Humphrey, R.R. (1987) 90 years and 535 miles: vegetation

changes along the Mexican border. University of New Mexico

Press, Albuquerque.

Humphrey, R.R. (1991) Montevideo Valley and its tallest

recorded cirio. Cactus and Succulent Journal, 63, 239–

240.

Humphrey, R.R. & Humphrey, A.B. (1990) Idria columnaris:

age as determined by growth rate. Desert Plants, 10,

51–54.

Hurd, W.E. (1929) Tropical cyclones of the eastern North

Pacific Ocean. Monthly Weather Review, 57, 43–49.

INEGI (2001) Sıntesis de informacion geografica del estado de

Baja California. Instituto Nacional de Estadıstica, Geografıa

e Informatica, Aguascalientes.

INE-SEMARNAP (1997) Programa de conservacion de la vida

silvestre y diversificacion productiva en el sector rural 1997–

2000. Instituto Nacional de Ecologıa, Secretarıa del Medio

Ambiente, Recursos Naturales y de Pesca, Mexico.

Jordan, P.W. & Nobel, P.S. (1981) Seedling establishment of

Ferocactus acanthodes in relation to drought. Ecology, 62,

901–906.

McAuliffe, J.R. (1988) Markovian dynamics of simple and

complex desert plant communities. American Naturalist,

131, 459–490.

McAuliffe, J.R. (1991) Demographic shifts and plant succes-

sion along a late Holocene soil chronosequence in the

Sonoran Desert of Baja California. Journal of Arid Environ-

ments, 20, 165–178.

McAuliffe, J.R. (1994) Landscape evolution, soil formation,

and ecological patterns and processes in Sonoran Desert

bajadas. Ecological Monographs, 64, 111–148.

Markham, C.G. (1972) Baja California’s climate. Weatherwise,

25, 64–76.

Martınez, P.L. (1965) Guia familiar de Baja California 1700–

1900. Editorial Baja California, Mexico.

Meko, D., Stockton, C.W. & Boggess, W.R. (1995) The tree-

ring record of severe sustained drought. Water Resources

Bulletin, 31, 789–801.

Midgley, J.J., Cowling, R.M., Hendricks, H., Desmet, P.G.,

Esler, K. & Rundel, P. (1997) Population ecology of tree

succulents (Aloe and Pachypodium) in the arid western

Cape: decline of keystone species. Biodiversity and Con-

servation, 6, 869–876.

Minnich, R.A. & Franco-Vizcaıno, E. (1998) Land of chamise

and pines: historical accounts and current status of northern

Baja California’s vegetation. University of California Publi-

cations in Botany, 80, 1–166.

Moran, R. (1962) Visitors to the flowers of Pachycereus

pringlei. National Cactus and Succulent Journal, 17, 21.

Moran, R. (1968) Cardon. Pacific Discovery, 21, 2–9.

Moriarty, J.R. (1968) Climatologic, ecologic and temporal

inferences from radiocarbon dates on archaeological sites,

Baja California, Mexico. Pacific Coast Archaeological Society

Quarterly, 4, 11–38.

Nabhan, G.P., Donovan, J., Buchmann, S.L. & Burquez, A.

(1999) Riqueza y estabilidad de los polinizadores de los

cirios (Fouquieria columnaris) del desierto sonorense: un

analisis comparativo. Boletın de la Sociedad Botanica de

Mexico, 65, 39–50.

Nabhan, G.P. & Suzan, H. (1994) Boundary effects on en-

dangered cacti and their nurse plants in and near a Sonoran

Desert biosphere reserve. Ironwood: an ecological and cul-

tural keystone of the Sonoran Desert (ed. by G.P. Nabhan and

J.L. Carr), pp. 55–67. Conservation International, Wash-

ington, DC.

Niering, W., Whittaker, R.H. & Lowe, C.H. (1963) The

saguaro: a population in relation to its environment. Science,

142, 15–23.

Niklas, K.J., Molina-Freaner, F. & Tinoco-Ojanguren, C.

(1999) Biomechanics of the columnar cactus Pachycereus

pringlei. American Journal of Botany, 86, 767–775.

Niklas, K.J., Molina-Freaner, F., Tinoco-Ojanguren, C. &

Paolillo, D.J., (2002) The biomechanics of Pachycereus

pringlei root systems. American Journal of Botany, 89, 12–21.

Nilsen, E.T., Sharifi, M.R., Rundel, P.W., Forseth, I.N. &

Ehleringer, J.R. (1990) Water relations of stem succu-

lent trees in north-central Baja California. Oecologia, 82,

299–303.

Parker, K.C. (1987) Site-related demographic patterns of organ

pipe cactus populations in southern Arizona. Bulletin of the

Torrey Botanical Club, 114, 149–155.

Parker, K.C. (1988) Growth rates of Stenocereus thurberi and

Lophocereus schottii in southern Arizona. Botanical Gazette,

149, 335–346.

Parker, K.C. (1993) Climatic effects on regeneration trends for

2 columnar cacti in the northern Sonoran Desert. Annals of

the Association of American Geographers, 83, 452–474.

Paulın-Ramırez, O., Navarro-Cordova, A., Morales-Guiza,

V.F., Rosinol-Monges, J.R., Preciado-Torres, A., Aguirre-

Wallace, C.A., Peralta-Castro, G. & Camacho-Camacho, J.A.

(1981) La determinacion regional de los coeficientes de agos-

tadero para el estado de Baja California. Secretaria de Agri-

cultura y Recursos Hidraulicos, Ensenada.

Penalba, M.C. & Van Devender, T.R. (1998) Cambios de

vegetacion y clima en Baja California, Mexico, durante los

ultimos 20,000 anos. Geologıa del Noroeste, 2, 21–23.

Pierson, E.A. & Turner, R.M. (1998) An 85-year study of

saguaro (Carnegiea gigantea) demography. Ecology, 79,

2676–2693.

Ramırez Apud Lopez, Z. (2002) Modelos de crecimiento para la

estimacion de la edad con aplicacion en el cirio (Fouquieria

columnaris). Master’s Thesis, Facultad de Ciencias, Univer-

sidad Autonoma de Baja California, Ensenada.

Rhode, D. (2002) Early Holocene juniper woodland and cha-

parral taxa in the central Baja California peninsula, Mexico.

Quaternary Research, 57, 102–108.

Rogers, G.F., Malde, H.E. & Turner, R.M. (1984) Bibliography

of repeat photography for evaluating landscape change.

University of Utah Press, Salt Lake City, UT.

Sankey, J.T., Van Devender, T.R. & Clark, W.H. (2001) Late

Holocene plants, Catavina, Baja California, Mexico. South-

western Naturalist, 46, 1–7.

S. H. Bullock et al.

142 Journal of Biogeography 32, 127–143, ª 2005 Blackwell Publishing Ltd

Schmidt, R.H. Jr (1989) The arid zones of Mexico: Climatic

extremes and conceptualization of the Sonoran Desert.

Journal of Arid Environments, 16, 241–256.

Serra Castelan, S. (1971) Hurricanes and tropical storms of the

west coast of Mexico. Monthly Weather Review, 99, 302–308.

Shreve, F. (1910) Rate of establishment of the giant cactus.

Plant World, 13, 235–240.

Shreve, F. (1935) The longevity of cacti. Cactus and Succulent

Journal, 7, 66–68.

Shreve, F. (1936) The transition from desert to chaparral in

Baja California. Madrono, 3, 257–264.

Shreve, F. (1942) The deserts of North America. Botanical

Review, 8, 195–246.

Shreve, F. (1964) Vegetation of the Sonoran Desert. Vegetation

and flora of the Sonoran Desert (ed. by F. Shreve and I.L.

Wiggins), pp. 9–186. Stanford University Press, Stanford, CA.

Southworth, J.R. (1889) El territorıo de la Baja California

Mexico ilustrada: Su agricultura, comercio, minerıa e indus-

trias en Ingles y Espanol. Reprint, 1989, Government of Baja

California Sur, La Paz.

Steenbergh, W.F. & Lowe, C.H. (1983) Ecology of the saguaro:

III. Growth and demography. National Park Service Scientific

Monograph Series 17, Washington, DC.

Thawley, J. (1968) Big bad Baja: the first annual Mexican 1000

rally. Hot Rod, 21, 46–47.

Thompson, R.S., Hostetler, S.W., Bartlein, P.J. & Anderson,

K.H. (1998) A strategy for assessing potential future changes

in climate, hydrology, and vegetation in the western United

States. US Geological Survey Circular, 1153, 1–20.

Tschirley, F.H. & Wagle, R.F. (1964) Growth rate and

population dynamics of jumping cholla (Opuntia fulgida

Engelm.). Journal of the Arizona Academy of Science, 3,

67–71.

Turner, R.M. (1990) Long-term vegetation change at a fully

protected Sonoran desert site. Ecology, 71, 464–477.

Turner, R.M., Bowers, J.E. & Burgess, T.L. (1995) Sonoran

Desert plants: an ecological atlas. University of Arizona Press,

Tucson, AZ.

Turner, R.M., Webb, R.H., Bowers, J.E. & Hastings, J.R. (2003)

The changing mile revisited. University of Arizona Press,

Tucson, AZ.

Van Devender, T.R. (1990) Late Quaternary vegetation and

climate of the Sonoran Desert, United States and Mexico.

Packrat middens: the last 40,000 years of biotic change (ed. by

J.L. Betancourt, T.R. Van Devender and P.S. Martin), pp.

134–163. University of Arizona Press, Tucson, AZ.

Vogel, J.C. (1974) The life span of the kokerboom. Aloe, 12,

66–68.

Watson, I.W., Westoby, M. & Holm, A.M. (1997) Continuous

and episodic components of demographic change in arid

zone shrubs: Models of two Eremophila species from Wes-

tern Australia compared with published data on other spe-

cies. Journal of Ecology, 85, 833–846.

Webb, R.H. & Betancourt, J.L. (1992) Climatic variability and

flood frequency of the Santa Cruz River, Pima County,

Arizona. US Geological Survey Water Supply Paper, 2379,

40 pp.

Webb, R.H. & Wilshire, H.G. (eds) (1983) Environmental ef-

fects of off-road vehicles. Springer-Verlag, New York.

Wells, P.V. (2000) Pleistocene macrofossil records of four-

needled Pinyon or Juniper encinal in the northern Vizcaıno

desert, Baja California del Norte. Madrono, 47, 189–194.

BIOSKETCHES

Stephen Bullock studies plant demography, phenology and

reproductive biology in arid and semi-arid ecosystems,

particularly on the Pacific slope of Mexico.

Nora Martijena is interested in relating population and

community dynamics of forests and arid scrublands to biotic

and abiotic factors, using survey, experimental and dendro-

chronological techniques.

Robert Webb studies landscape change in the Sonoran,

Mojave and Great Basin Deserts, emphasizing historical

changes in upland vegetation, vegetation recovery and the

effects of landscape changes on fluvial systems.

Raymond Turner works on phytogeography, plant ecology

and ecosystem management; his early interest in vegetation

dynamics prompted establishment of long-term vegetation

study plots as well as the use of repeat photography for docu-

menting landscape change in the Sonoran Desert and Kenya.

Demographic changes in cirio and cardon

Journal of Biogeography 32, 127–143, ª 2005 Blackwell Publishing Ltd 143