ALLOZYMIC VARIATION IN DESERT PUPFISH FROM NATURAL AND
ARTIFICIAL HABITATS: GENETIC CONSERVATION IN
FLUCTUATING POPULATIONS
Jason B. Dunham'
and
W.L. Minckley
Department of Zoology. Arizona State University. Tempe, AZ
85287-1501
Keywords: Conservation genetics, desert pupfish, Cyprinodon
macularius, Colorado River, Salton Sea.
First running head: J.B. Dunham & W.L. Minckley
Second running Head: Conservation genetics of an endangered
fish
'Present address:
Biology Department and Program in
Ecology. Evolution and Conservation Biology. University of
Nevada. Reno. NV 89557
2
ABSTRACT
Few if any consistent differences in patterns of
allozymic variation exist between and among wild and captive
populations of endangered desert pupfish (Cyprinodon m.
macularius). Those which do can be attributed to different
founder sources, founder and bottleneck effects or a
combination of all three. Local isolation, natural or
artificial, promotes differentiation in both wild and
captive populations. This was countered over the natural
long-term condition of drought by an alternative state of
dispersal and population mixing during wetter periods. Such
gene flow among populations is now precluded by hydrologic
control. Recovery of desert pupfish to self-sufficiency
will require re-establishment of a natural hydrograph, which
for the foreseeable future is unlikely in the lower Colorado
River basin. If this charismatic desert fish is to survive,
resource managers must combine knowledge of ecological and
genetic patterns with political realities in a captive
management program involving periodic isolation followed by
production of a composite population that is isolated then
again combined, repeatedly, until natural conditions are reestablished.
3
INTRODUCTION
Preservation of biological diversity at the most basal level
entails maintenance of genetic variation (Frankel & Soule,
1981; Schonewald-Cox et al., 1983). Recovery of imperiled
taxa requires genetic variation be conserved to avoid
deleterious effects of inbreeding and sustain their
evolutionary potential. Differing patterns in magnitude and
distribution of genetic variation among taxa demand
different conservation measures (Templeton, 1991; Templeton
et al., 1991). As a result, biochemical surveys of genetic
variation have become routine parts of the recovery process
(Hedrick & Miller, 1992).
One drawback of such surveys is the uncertainty of
quantifying effects of factors contributing to observed
patterns (Rockwell & Barrowclough, 1987). For designing a
conservation plan, genetic variation must be interpreted in
light of ecological and demographic information.
Unfortunately, while surveys of genetic variation have
become relatively simple, ecological and demographic data
are often unavailable and difficult to obtain (Stacey &
Taper, 1992). Studies of translocated populations with
known histories may thus prove particularly useful in
interpreting patterns observed in the field (Easteal, 1981).
4
Recovery efforts for imperiled desert fishes in North
America frequently involve translocations (Minckley &
Brooks, 1985; Hendrickson & Brooks, 1991; Minckley, 1995).
Translocations can reduce short-term extinction
probabilities by increasing total numbers of individuals and
populations, but such efforts are arguably effective in
maintaining the genetic integrity of imperiled taxa
(Hendrickson & Brooks, 1991). To be ultimately successful,
translocations must preserve genetic variation necessary for
a taxon's persistence over evolutionary time. Studies of
genetic variation in translocated fishes have helped
identify important population patterns and processes and
provide an important basis for directing future recovery
efforts (see Echelle, 1991 and references therein; DeMarais
& Minckley, 1993; Stockwell, 1995).
This paper presents results of a study of allozymic
variation in wild and captive populations of desert pupfish
(Cyprinodontidae, Cyprinodon macularius Baird & Girard), an
endangered fish native to the lower Colorado River basin
(Fig. 1). Here the term "captive" is used to refer to
translocated desert pupfish populations in quasi-natural
outdoor habitats including fish hatcheries, concrete ponds
and other human-made environments such as irrigation ponds
and excavated natural springs. Hendrickson & Brooks (1991)
defined translocated (transplanted) populations as
those created by human transport of fishes from one locality
5
to another, but restricted their analysis to "wild"
habitats. There is undoubtedly some overlap between our
definitions of "wild" habitats because pupfish are capable
of maintaining self-sustaining populations in a wide variety
of conditions.
Genetic variation among captive and wild populations
was surveyed to address two questions. Do patterns of
genetic variation differ among wild and captive populations?
If so, what factors are responsible? Patterns in captive
populations are examined in light of number of founders,
time since founding, patterns of gene flow and relative
habitat sizes. Variation among wild populations, which have
patterns similar to those in captivity, are explained in the
context of present and historic habitat conditions.
Episodic interbreeding then isolation, bottlenecking and reconnection, followed again by interbreeding, comprise the
processes resulting in the inferred and genetic patterns. A
rationale is developed for conservation based on hands-on
manipulation of captive metapopulations.
MATERIALS AND METHODS
Study organism
Desert pupfish was formerly distributed throughout much
of the vast lower Colorado River basin of northwestern
6
Mexico (Baja California Norte, Sonora) and southwestern
United States (Arizona, California) (Fig. 1). Two
subspecies are recognized, C. m. eremus Miller & Fuiman,
from Quitobaquito Spring, Organ Pipe Cactus National
Monument, Arizona and C. m. macularius Baird & Girard in
most of the remainder of its geographic range. A third form
in the Rio Sonoyta, Sonora is not yet named (Miller &
Fuiman, 1987).
Only C. m. macularius is considered here.
This taxon is part of a declining native fish fauna in
the American Southwest (Minckley & Deacon, 1991). Habitat
destruction and impacts of exotic species have extirpated
and fragmented populations throughout its range (Schoenherr,
1988; Minckley et al., 1991), and it was listed as
endangered by the U.S. Fish & Wildlife Service (USFWS,
1986).
A recovery plan (Marsh & Sada, 1994) included
provisions for maintaining genetic variation, some of which
are incorporated here.
Sampling localities and collections
Pupfish were obtained from nine captive populations:
Dexter National Fish Hatchery and Technology Center
(hereafter Dexter), Chavez County, New Mexico; private pond
(W.L. Minckley residence; WLM) and Desert Botanical Garden,
Maricopa County, Arizona; Boyce-Thompson Arboretum, Pinal
County, Arizona; Howard Well, Graham County, Arizona;
Flowing Wells Junior High School, Pima County, Arizona;
7
Simone/McCallum Pond and Living Desert Reserve, Riverside
County, California; and Palm Canyon Pond, San Diego County,
California. Wild fish were obtained from four sites around
the Salton Sink (Imperial and Riverside counties,
California): Salt Creek, Cleveland Street Drain, unnamed
shoreline pool
-
1.0 km east of Cleveland Street Drain, and
San Felipe Creek (Fig. 1). Fish were frozen at capture and
stored at -80°C or transported and maintained alive until
processed.
Historical data on captive populations (Fig. 2) are
from unpublished notes of Brian Bagley (formerly Arizona
Game and Fish Department), Betsy Bolster (California
Department of Fish and Game), Gerald Burton (formerly
USFWS), and Paul C. Marsh and W.L. Minckley (Arizona State
University; ASU). Captive fish in Arizona and New Mexico
were of two separate lineages from Sonora: Canal Sanchez
Taboada (Dexter, Desert Botanical Garden), and "Terrace
Springs" on the Colorado River delta (WLM, Howard Well,
Flowing Wells). Boyce-Thompson fish were a mixture of both
Mexican sources. California captive populations each were
from a different wild founder population: Living Desert from
the Salton Sea, Palm Canyon from San Felipe Creek and
Simone/McCallum from Salt Creek.
Electrophoretic techniques
Living fish were euthanized with chlorobutanol before
dissection. Water-soluble protein extracts were obtained
from eye, muscle and liver. Tissues were placed in 1.5-ml
microcentrifuge tubes, homogenized in equal volumes of
distilled water and centrifuged for 5 min at
-
18,000 G
(Murphy et al. 1990). Supernatants were loaded on filterpaper wicks, inserted into 12W starch gels, electrophoresed
for -12 hr at 25-50 mA then sliced and stained for enzyme
activity. Proteins, tissue distributions and buffers are in
Table 1. Proteins surveyed were selected from those
previously reported polymorphic (p<0.99) in C. macularius (4
of 6 loci recorded by Turner [1983, 1984]) and other
pupfishes (Echelle et al., 1987). Other criteria included
cost and ease of resolution.
Data analyses
Data analyses used the BIOSYS-1 computer program of
Swofford & Selander (1981). Loci were deemed polymorphic if
frequency of the most common allele was <0.99. Average
heterozygosity per locus per individual was calculated
following Nei (1978). Note that estimates of heterozygosity
are inflated since only polymorphic loci were included. The
chi-square goodness-of-fit test with Levene's (1949)
correction was used to test for deviations from HardyWeinberg expectations. Significance of heterogeneity among
9
samples was determined by the heterogeneity chi-square
statistic. Wright's (1978) F-statistics were applied to
determine hierarchical distribution of genetic variation
among populations.
For hierarchical analysis, populations were arranged as
follows (Figs. 1, 2): lower Colorado delta [(Dexter, Desert
Botanical Garden) (WLM, Howard Well, Flowing Wells) (BoyceThompson)], Salton Sink [(Salt Creek, Simone/McCallum)
(Cleveland drain) (shorepool) (San Felipe Creek, Palm
Canyon) (Living Desert)]. Genetic variation was thus
partitioned into the proportion accounted for within sites,
among sites within lineages (within parentheses), among
lineages within areas (enclosed by brackets, lineages
defined as a group of populations descended from a common
source) and among areas (lower Colorado Delta and Salton
Sink, between brackets).
Observed patterns in delta captives were analyzed
relative to known population variables using procedures of
Mantel (1967). A "two-matrix" Mantel test (Douglas &
Endler, 1982; Manly, 1992) was used to compute association
between distance matrices of variation (heterozygosity) and
four variables: number of founders, relative habitat size,
time since founding, and hierarchical level. The last
reflects the number of known founding events in a
population's history, e.g. captives at Dexter and WLM were
10
separated from nature by a single founder event each, while
those at Flowing Wells had experienced four such events.
Captives derived from sources of differing
heterozygosity may differ genetically for that reason alone.
To control for this potentially confounding factor, we used
the standardized difference between source and descendant
g„, defined as follows:
population heterozygosity,
HDIFF
=
I
4-1 -11x I /11X-1
1
( 1)
where H is the source population heterozygosity and 14, is
descendant population heterozygosity. Mantel tests for
association with number of founders, relative habitat size
and time since founding used
RinFF
as the response variable.
The difference between original source and descendant
population heterozygosity
( HsouRcE)
was used as the response
variable to determine the effect of hierarchical level.
HsOURCE
is defined as follows:
HSOURCE =
1 H0 -
H
I /H0
(2)
where H, is the original source heterozygosity and x the
number of founding events separating the two populations.
For example, the captive source of Flowing Wells was WLM,
and they are separated by three founder events (Fig. 2).
Gaps in the history of pupfish populations are many
(Fig. 2) and potentially problematic. For example, the
captive source of Flowing Wells was extirpated by the time
of this study so it was compared to its nearest population,
Howard Well. One population, Boyce-Thompson, was founded
11
from a mixture of two captive sources (Fig. 2) so
heterozygosity values from each source were averaged for
comparisons. The same is true for the source of Living
Desert pupfish. In cases such as the desert pupfish, where
critical information is incomplete or possibly biased, the
most pragmatic approach is to analyze the data first and
then carefully examine results for discrepancies due to such
inadequacies (Elliot, 1992).
RESULTS AND DISCUSSION
Overall, four polymorphic loci were resolved by
procedures outlined in Echelle et al. (1987) and Turner
(1983, 1984); this is consistent with generally low levels
of allozymic variation in the genus Cyprinodon (Echelle
1991). Statistically significant deviation from HardyWeinberg expectations occurred in only 2 of 39 chi-square
tests, an excess of heterozygotes for sAh-A at BoyceThompson and a deficiency of heterozygotes for Gpi-A at
Simone/McCallum (Table 2). Alleles per locus, polymorphic
loci and heterozygosities are in Table 3.
Desert pupfish populations exhibited considerable
genetic variation at these loci (Tables 2 and 3). Captive
populations originating on the delta were fixed for the same
allele at the Gpi-B locus that dominated in Salton Sink
populations (>90% frequency in 6 of 7 populations and fixed
12
in Living Desert). Among delta-derived populations, Dexter
and Desert Botanical Garden fish lacked variation at the
Gpi-A and Ldh-A loci. Two populations (Howard Well, Flowing
Wells) showed marked differences in allele frequency at the
sAh-A locus. Variation was less among Salton Sink samples,
but some differences in frequencies were nonetheless
apparent (e.g. sAh-A and Ldh-A at Living Desert).
Populations were clustered into four lineages (=captive
populations with common ancestry) to more specifically
examine divergence. Chi-square heterogeneity (Table 4)
showed significant frequency divergences in three of the
four. Hierarchical analysis of variation (Table 5)
indicated 88.5*
(100-100FsITE-ToTAL)
of total genic diversity is
contained in an average site. If partitioned by geographic
area, 86.3* is contained in an average site within a
geographic area ( 100 -100F„TE -AREA)
•
Note that FAREA _ TOTAL is
slightly negative, which explains FSITETOTAL being <F„,"-AREA
(Table 5).
Genetic variation among captive populations was
strongly associated with founder population size, but not
time since founding, relative habitat size or hierarchical
relationships. The Mantel test indicated a strong negative
association (P=0.016) between number of founders and
difference in heterozygosity between source and descendant
populations (11,„„„).
This result is consistent with
theoretical expectations for species that can rebound
13
quickly following a founding event or population bottleneck
(Nei et al. 1975). This may also explain why time since
founding has little influence on retention of genetic
variability in captive populations.
Lack of association between genetic variability and
relative habitat size may be explained by the highly
variable nature of pupfish habitat. Habitat instability,
especially fluctuations in water levels, can lead to
dramatic changes in population size. This is certainly the
case with Howard Well and Boyce-Thompson pupfish (B. Bagley,
pers. comm. 1990). Such variations may obscure the impact
of multiple founding events (hierarchical level) as well.
Levels of genetic variation in the original wild
sources of captive pupfish populations are unknown and form
an important and potentially confounding gap in the
historical data. Captive populations derived from sources
with high heterozygosity should be more heterozygous than
those derived from less variable sources. Overall, source
and captive population variation are not significantly
related (Mantel test P>0.05), but there are no data on
genetic variation of source populations at the times of
initial collections. If significant genetic changes in
source populations occurred after captives were founded,
present genetic patterns may be invalid for comparisons.
Observed allele frequencies in Dexter and WLM captive
populations point strongly toward the influence of source
14
populations on genetic variation. The Dexter population was
founded with 280 fish in 1983 and subsequently maintained as
a few thousand individuals. The WLM population originated
with only 64 fish in 1976 and persisted in the low hundreds.
Founder effects should thus be more pronounced in the WLM
population and less in Dexter had the sources been the same,
exactly opposite the lack of variation in Dexter relative to
the originally and chronically smaller and older WLM
population (Table 2). Sources were both from the same part
of the Colorado Delta, yet the pattern of variation
indicates differentiation despite geographic proximity.
Wild founders for Dexter were from a pool excavated by
heavy equipment near the terminus of Canal Sanchez Taboada
at Santa Clara Slough (G. Burton, pers. comm.). The latter
is a variably-sized marsh on the southeastern edge of the
delta, now partially maintained by artificial inflow from
the canal. Low genetic variation in Dexter is explained if
the wild founder population either originated from a few
individuals, suffered a severe bottleneck(s) after
establishment, or both. Founders of the WLM population were
from a series of five springheads (collectively flowing an
estimated <15 L/min) rising a few tens of meters apart along
the base of an alluvial terrace, almost level with and
flowing toward the adjacent Santa Clara Slough (Fig. 1).
They thus may have included individuals from as many as five
separate populations, none of which likely exceeded -100t200
15
fish, and greater variability might be anticipated. Natural
habitats like these springs may, however, appeai isolated at
a point in time, when in fact they are periodically
interconnected. Nearby vegetation was inundation tolerant
(desert plants were absent) indicating periodic flooding. A
rise in water level of <1.0 m later inundated four and
perhaps all five springs, allowing gene flow. Genetic
patterns among captives thus indirectly suggest a degree of
genetic fragmentation of wild populations on the Colorado
Delta, consistent with the fragmented condition of habitats
(owing to desiccation) at the time the wild collections were
made (Hendrickson & Varela, 1989).
In contrast to delta pupfish, wild populations in the
Salton Sink are relatively homogeneous over a much larger
geographic scale. Prior to 1904, pupfish populations in the
sink were restricted to isolated springs and seeps (Hubbs &
Miller, 1948). In 1904-1907, flooding on the Colorado River
entered to form the present Salton Sea (Sykes, 1937), and
pupfish soon formed remarkably large shoreline populations
(Coleman, 1929; Walker, 1961; Barlow, 1958; 1961).
There is
little doubt that numerous, formerly isolated local
populations contributed to this expansion. The species
remained abundant until the mid-1960's, followed by dramatic
declines in apparent negative response to introduction and
establishment of exotic fishes (Schoenherr, 1988).
16
More recently, pupfish populations have rebounded
following decimation of exotic fishes in the Salton Sea due
to increasing salinity accumulated through irrigation return
flow and evaporation (B. Bolster, A. Schoenherr, pers. comm.
1992). Thus, over the past century, desert pupfish
populations in the Salton Sink have dramatically expanded
and contracted in response to both natural and artificial
limiting factors.
A similar "boom and bust" cycle occurred on the
hydrologically connected Colorado Delta, where pupfish
populations must have historically waxed and waned with vast
fluctuations in annual discharge of the Colorado River (Graf
1985). Fish endured increasing desiccation following
upstream agricultural diversions beginning in the early
1900's (Sykes, 1914) and regulation of river flows beginning
with Hoover (Boulder) Dam in 1935 (Graf, 1985). Rewatering
of the delta by extensive flooding in 1983-84 resulted in
increased distribution and abundance of pupfish (Hendrickson
& Varela, 1989) which persists in part today (Abarca et al.,
1992). The genetic impacts of this population surge
following rewatering are unknown.
CONCLUSIONS AND MANAGEMENT IMPLICATIONS
We conclude, as did Turner (1983, 1984), that desert
pupfish fail to demonstrate consistent patterns of
17
interpopulation variation. However, genetic divergence
among populations is not uncommon. It appears that drift
(manifested via founder/bottleneck effects) is the most
important determinant of variation in captive populations.
Genetic variation among sites is greater than among
lineages, which is, in turn, greater than variation between
geographic areas. Such a pattern points to potentially
strong influences of within-population events in determining
genetic variation of captive populations.
In terms of current models of metapopulation structure,
desert pupfish are most closely approximated by mainlandisland models (Harrison, 1991). Several "mainland" areas
such as permanent spring and stream refugia are (or were)
periodically connected by flooding (Sykes, 1937; Hendrickson
& Varela, 1989; Miller et al., 1991). Intermittently
flooded areas also harbor flourishing populations until
eventual desiccation once again isolates more permanent
habitats.
Such a system can persist in hot deserts only if water
remains available and reliable refugia and flooding are
frequent enough to ensure regional survival and periodic
gene flow among populations. Original geographic range of
desert pupfish exceeded 40,000 km2 over an altitudinal range
of -50 to 1500 m (Minckley, 1973). Upstream lay the seventh
largest watershed in the coterminous USA, part of which
yielded at least some surface water each year (Minckley,
18
1991a).
Faulting and other phenomena associated with active
tectonism created fracture zones where springs were
concentrated (Minckley et al., 1986). Large snowpack in
unusually wet years, and monsoonal rains in others assured
periodic floods. In this situation the pupfish enjoyed a
high probability of survival.
The prognosis is poor for continued survival of wild
desert pupfish populations. With disappearance from the
entire Gila River basin (Fig. 1), an unknown but substantial
proportion of intraspecific diversity may have been lost.
Perhaps more importantly, human hydrological control over
the rivers precludes gene flow among populations. Desert
pupfish will likely not achieve self sufficiency until
natural hydrologic conditions, including major flooding, are
restored. Considering demands on Colorado River water by
more than 20 million people (Graf, 1985), such seems
unlikely in the foreseeable future. Hands-on management
of captive populations that mimics a more natural state
should therefore be implemented.
Present recovery efforts for desert pupfish focus on
maintaining local captive populations, which is inconsistent
with natural history of the species. If numbers of local
populations are or become small, persistent isolation will
result in a situation corresponding to the "Death Valley
Model" of Meffe and Vrijenhoek (1988). Theoretical genetic
models predict a decline in both among- and within-
19
population genetic diversity in such cases (Chakraborty
& Leimar, 1987), clearly an undesirable result.
An alternative plan (Fig. 3) represents a compromise
between logistical and biological considerations. Existing
captive populations founded directly from wild sources are
termed "primary populations," which should be periodically
composited to form a pool of "migrants." The composited
population could also be used as a source to create (and
serve as a source of migrants for) additional captive
populations. These are termed "secondary populations"
because they were not directly founded with wild
individuals. When available, wild fish should be added to
their descendant primary populations, thus ensuring captive
populations retain their original genetic integrity.
Strict adherence to the plan outlined in Figure 3 is
less critical than recognizing the need for establishing and
maintaining more captive populations with a degree of amongpopulation gene flow. Specific recommendations regarding
numbers and sizes of captive populations and amongpopulation migration rates must await future study. For
this reason, it is strongly recommended that future recovery
efforts be implemented on an experimental basis.
Since the small number of captive populations may not
now include all genetic materials upon which the species'
well-being depends, larger numbers of and diversities in
captive populations are desirable. Searches for and study
20
of wild populations should therefore continue for a time, to
minimize the likelihood that "evolutionary significant
units" (Waples, 1991) are not overlooked. Furthermore, the
possibility of local adaptation (e.g. Hirshfield et al.,
1980) should be considered before groups of wild pupfish
populations are managed as relatively homogeneous units.
Efforts toward perpetuation of desert pupfish as a
single, diverse genome should, however, commence as soon as
a decision is made that no new variability nor wild
populations are forthcoming. The diverse provenance,
histories and even genetic differences among existing
captive populations all speak to relatively high collective
variability. They are no more or less than isolates from an
apparently common genome available for recolonization of
rejuvenated habitats. Observed patterns of variation in
typically small, remnant populations have little to do with
whether their origins and habitats are natural or artificial
(see also Turner, 1984). The captives thus differ little if
at all from wild pupfish that must have responded in the
distant past and in the 1980's when conditions improved
after isolation by drought on the Colorado Delta, in springs
on the desiccated floor of the Salton Sink in the period
1904-07 or around the Salton Sea after the return of
favorable biotic conditions in the late 1980's.
The severe selective arenas occupied by desert pupfish
are not difficult to replicate. The fish needs little more
21
than water to survive so long as other species are excluded.
Fortunately, most non-native species, considered critical
problems in the conservation of most native southwestern
species (Minckley, 1991h), cannot survive the severe
conditions pupfish accommodate readily.
Among ecological
and genetic stresses pupfish have experienced for millennia
isolation, sudden expansions and reductions in habitat size
or quality and mixing and interbreeding of populations,
among others, none is difficult nor expensive to provide.
In fact, some comparable "manipulations" inadvertently occur
through human error during refuge operations (Minckley et
al., 1991; Minckley, 1995). Perhaps these past "mistakes"
can be applied to favor the species.
The desert pupfish has become somewhat of a standard
bearer for imperiled fishes of western North America, with
charismatic status in the public eye as a representative for
desert aquatic ecosystems. Its perpetuation should be of
high priority.
22
ACKNOWLEDGEMENTS
Thanks to T. Dowling, P. Marsh and M. Douglas for their
invaluable assistance and perseverance; A.A. Echelle and
A.E. Echelle for their valuable advice and hospitality; M.
Brown, M. Cardenas, B. DeMarais, K. Nichol, A. Schoenherr,
R. Timmons, C. Williams for help with field collections; P.
Hedrick and M. Childs for comments on an early draft, and B.
Bolster for help with collecting permits and funding.
Partial financial support was provided by California
Department of Fish and Game, American Museum of Natural
History Theodore Roosevelt Fund and Sigma Xi Grants-in-Aid,
and Arizona State University Department of Zoology. Pupfish
collections were authorized by US Fish & Wildlife Service
permits 702631 (Region I) and PRT-676811 (Region II) in
cooperation with California Department of Fish and Game and
Arizona Game and Fish Department, respectively.
REFERENCES
Abarca, F.J., D. Weedman, & A. Varela-Romero. 1992. Fish.
Pages 18-20 in: Field Report 1. Study of cienega Santa
Clara and environs. Report to U.S. Fish and Wildlife
Service, Albuquerque, New Mexico. Cienega Santa Clara
Working Group (available through Nongame Branch, Ariz.
Game Fish Dep., Phoenix, AZ). 34pp, 6 maps.
Barlow, G.W. 1958. Daily movements of desert pupfish,
Cyprinodon macularius, in shore pools of the Salton
Sea, California. Ecol., 39, 580-87.
Barlow, G.W. 1961. Social behavior of the desert pupfish,
Cyprinodon macularius, in the field and in the
aquarium. Am. Midi. Nat., 65, 339-58.
Buth, D.G. 1983. Duplicate isozyme loci in fishes:
origins, distribution, phyletic consequences and locus
nomenclature. In Current topics in biological and
medical research, Volume 10, ed. M.C. Rattazzi, J.G.
Scandalios & G.S. Whitt. Liss, New York, pp. 381-400.
Chakraborty, R. & D. Leimar. 1987. Genetic variation
within a subdivided population. In Population genetics
and fisheries management, ed. N. Ryman & F. Utter.
Univ. Wash. Press, Seattle, pp. 89-120.
Coleman, G.A. 1929. A biological survey of the Salton Sea.
Calif. Dep. Fish Game Bull., 15, 218-27.
24
DeMarais, B.D. & W.L. Minckley. 1993. Genetics and
morphology of Yaqui chub Gila purpurea, an endangered
cyprinid fish subject to recovery efforts. Biol.
Conserv., 66, 195-206.
Douglas, M.E. & J.A. Endler. 1982. Quantitative matrix
comparisons in ecological and evolutionary
investigations.
a.
Theor. Biol., 99, 777-95.
Easteal, S. 1981. The history of introductions of Bufo
marinus (Amphibia: Anura); a natural experiment in
evolution. Biol. 1. Linn. Soc., 16, 93-113.
Echelle, A.A. 1991. Conservation genetics and genetic
diversity in freshwater fishes of western North
America. In Battle against extinction: native fish
management in the American West, ed. W.L. Minckley &
J.E. Deacon. Univ. Ariz. Press, Tucson, pp. 141-53.
Echelle, A.A., A.F. Echelle & D.R. Edds. 1987. Population
structure of four pupfish species (Cyprinodontidae:
Cyprinodon) from the Chihuahuan desert region of New
Mexico and Texas: allozymic variation. Copeia, 1987,
668-81.
Elliot, J.M. 1992. Variation in the population density of
adult sea-trout, Salmo trutta, in 67 rivers in
England and Wales. Ecol. Fresh. Fish, 1, 5-11.
Frankel, O.H. & M.E. Soule. 1981. Conservation and
evolution. Cambridge University Press, Cambridge.
25
Graf, W.L. 1985. The Colorado River, instability and
management. Res. Publ. Geogr., Association of
American Geographers, Washington, D.C.
Harrison, S. 1991. Local extinction in a metapopulation
context: an empirical evaluation. Biol. J. Linn. Soc.,
42, 73-88.
Hedrick, P.W. & P.S. Miller. 1992. Conservation genetics:
techniques and fundamentals. Ecol. Appl., 2, 30-46.
Hendrickson, D.A. & J.E. Brooks. 1991. Transplanting
short-lived fishes in North American deserts: review
assessment and recommendations. In Battle against
extinction: native fish management in the American
West, ed. W.L. Minckley & J.E. Deacon. Univ. Ariz.
Press, Tucson, pp. 283-98.
Hendrickson, D.A. & A. Varela-Romero. 1989. Conservation
status of desert pupfish, Cyprinodon macularius, in
Mexico and Arizona. Copeia, 1989, 478-83.
Hirshfield, M.F., C.R. Feldmeth & D.L. Soltz. 1980.
Genetic differences in physiological tolerances of
Amargosa pupfish (Cyprinodon nevadensis) populations.
Science, 207, 999-1000.
26
Hubbs, C.L. & R.R. Miller. 1948. The zoological evidence:
correlation between fish distribution and
hydrographic history of the desert basins of western
United States. In The Great Basin with emphasis on
glacial and postglacial times. Bull. Univ. Utah, 38,
Biol. Ser. 10.
International Union of Biochemistry, Nomenclature Committee.
1984. Academic Press, Orlando, Florida.
Levene, H. 1949. On a matching problem arising in genetics.
Ann. Math. Stat., 20, 91-94.
Manly, B.F.J. 1992. Randomization and Monte Carlo methods
in
biology. Chapman and Hall, New York.
Mantel, N. 1967. The detection of disease clustering and a
generalized regression approach. Canc. Res., 27, 20920.
Marsh, P.C. & D.W. Sada. 1994. Desert pupfish, Cyprinodon
macularius, recovery plan. U.S. Fish Wildl. Serv., Reg.
2, Albuquerque, New Mexico.
Meffe, G.K. & R.C. Vrijenhoek. 1988. Conservation genetics
in the management of desert fishes. Conserv. Biol.,
2, 157-69.
Miller, R.R., C. Hubbs & F.H. Miller. 1991. Ichthyological
exploration of the American West: The Hubbs-Miller era,
1915-1950. In Battle against extinction: native fish
management in the American West, ed. W.L. Minckley &
J.E. Deacon. Univ. Ariz. Press, Tucson, pp. 19-42.
27
Miller, R.R. & L.A. Fuiman. 1987. Description and
conservation status of Cyprinodon macularius eremus,
a new subspecies of pupfish from Organ Pipe Cactus
National Monument, Arizona. Copeia, 1987, 593-609.
Minckley, W.L. 1973. Fishes 2f Arizona. Ariz. Game Fish
Dep., Phoenix.
Minckley, W.L. 1991.
A Native fishes of the Grand Canyon
region: an obituary? In Colorado river ecology and dam
management, Proceedings of a symposium May 24-25, 1990,
Santa Fe, New Mexico. National Academy of Science
Press, Washington D.C., pp. 124-77.
Minckley, W.L. 1991h. In J.N. Rinne (photographs) & W.L.
Minckley (text), Native fishes in arid lands: a
dwindling resource of the desert Southwest. Gen. Tech.
Rep. RN-206. Ft. Collins, Colorado: U.S. Dep. of
Agric., For. Ser., Rocky Mtn. For. Range Exp. Sta.
45 p.
Minckley, W.L. 1995. Translocation as a tool for conserving
imperiled fishes: experiences in western United States.
Biol. Conserv. 72: 297-309.
Minckley, W.L. & J.E. Brooks. 1985. Transplantations of
native Arizona fishes: records through 1980. J.
Ariz.-Nev. Acad. Sci., 20, 73-89.
Minckley, W.L. & J.E. Deacon, eds. 1991. Battle against
extinction: native fish management in the American
West. Univ. Ariz. Press, Tucson.
28
Minckley, W.L., D.A. Hendrickson & C.E. Bond. 1986.
Geography of western North American freshwater
fishes: description and relationships to
intracontinental tectonism. In Zoogeography at North
American freshwater fishes, ed. C.H. Hocutt & E.O.
Wiley. John Wiley and Sons, New York, pp. 519-613.
Minckley, W.L., Meffe, G.K. & D.L. Soltz. 1991. Conservation
and management of short-lived desert fishes: the
cyprinodontoids. In Battle against extinction: native
fish management in the American West, ed. W.L. Minckley
& J.E. Deacon. Univ. Ariz. Press, Tucson, pp. 247-82.
Murphy, R.W., J.W. Sites jr., D.G. Buth & C.H. Haufler.
1990. Proteins I: isozyme electrophoresis. In Molecular
systematics, ed. D.M. Hillis & C. Moritz. Sinauer
Assoc. Inc., Chicago, pp. 45-126.
Nei, M. 1978. Estimation of average heterozygosity and
genetic distance from a small number of individuals.
Genetics, 89, 583-90.
Nei, M., T. Maruyama & R. Chakraborty. 1975. The bottleneck
effect and genetic variability in populations.
Evolution, 29, 1-10.
Rockwell, R.F. & G.F. Barrowclough. 1987. Gene flow and
the genetic structure of populations. In Avian
genetics: A population and ecological approach, ed. F.
Cooke & P.A. Buckley. Academic Press, Orlando, Florida,
pp. 223-52.
29
Schoenherr, A.A. 1988. A review of the life history and
status of the desert pupfish, Cyprinodon macularius.
Bull. South. Cal. Acad. Sci., 87, 104-34.
Schonewald-Cox, C.M., S.M. Chambers, B. MacBryde & W.L.
Thomas, eds. 1983. Genetics and conservation.
Benjamin Cummings, Menlo Park, California.
Stacey, P.B., & M. Taper. 1992. Environmental variation
and the persistence of small populations. Ecol. Appl.,
2, 18-29.
Stockwell, C.A. 1995. Evolutionary trajectories in
recently established populations of western
mosquitofish (Gambusia affinis). Doctoral dissertation,
Univ. Nevada, Reno, 105 pp.
Swofford, D.L. & R.B. Selander. 1981. BIOSYS-1: A FORTRAN
program for the comprehensive analysis of
electrophoretic data in population genetics and
systematics. a. Hered., 72, 281-83.
Sykes, G. 1914. Geographical features of the Cahuilla Basin.
In The Salton Sea, ed. D.T. MacDougal. Carnegie Inst.
Washington Publ., 193, 13-20.
Sykes, G. 1937. The Colorado Delta. Carnegie Institute of
Washington Publ., 460, 1-193.
Templeton, A.R. 1991. The role of genetics in captive
breeding and reintroduction for species conservation.
Endangered Species UPDATE, 8, 14-17.
30
Templeton, A.R., K. Shaw, E. Routman, & S.K. Davis. 1991.
The genetic consequences of habitat fragmentation. Ann.
MQ. Bat. Gard., 77, 13-27.
Turner, B.J. 1983. Genic variation and differentiation of
remnant natural populations of the desert pupfish,
Cyprinodon macularius. Evolution, 37, 690-700.
Turner, B.J. 1984. Evolutionary genetics of artificial
refugium populations of an endangered species, the
desert pupfish. Copeia, 1984, 364-370.
U.S. Fish and Wildlife Service. 1986. Endangered and
threatened wildlife and plants; determination of
endangered status and critical habitat for the desert
pupfish. Fed. Regist., 51, 10842-10851.
Walker, B.W. 1961. The ecology of the Salton Sea,
California, in relation to the sportfishery. Calif.
Dep. Fish Game Bull., 113, 1-203.
Waples, R.S. 1991. Pacific salmon Oncorhynchus spp., and the
definition of "species" under the Endangered Species
Act. Mar. Fish. Rev., 53, 11-22.
Wright, S. 1978. Evolution and the genetics DI
populations, Volume 4. Variability within and among
natural populations. University of Chicago Press,
Chicago.
31
Table 1.
Presumptive loci, proteins, tissue sources, and
buffer systems surveyed in Cyprinodon m. macularius.
Enzyme
nomenclature follows recommendations of the International
Union of biochemistry (1984).
Locus designations follow
Buth (1983).
Tissue
Buffer
Scored
Systeml
sAh-A
liver
2
isomerase
Gpi-A
muscle-eye
1
(5.3.1.9)
Gpi-B
muscle-eye
1
L-Lactate dehydrogenase
Ldh-A
muscle-eye
1
Protein
(EC number)
Locus
Aconitase hydratase
(4.2.1.3)
Glucose-6-phosphate
(1.1.1.27)
Buffer systems as follows: 1) electrode buffer: 0.04 M
1
citric acid monohydrate, adjust to pH 6.1 by adding 10-15
mL/liter of n-3-aminopropylmorpholine; gel buffer: 1 in 19
dilution of electrode buffer. 2) electrode buffer: 0.1 M
tris hydroxymethylaminomethane (= "tris"), 0.3 M citric
acid, pH 7.5; gel buffer: 1 volume of electrode solution + 6
volumes HA.
32
Table 2. Allelic variation at four variable loci in 13
populations of Cyprinodon m. macularius. Numbers in
parentheses indicate sample sizes.
Population Locus: sAh-A
Gpi-A Gpi-B Ldh-A
San Felipe Creek
(29)
(29)
(29)
a: 0.172
0.966
0.931
0.052
b: 0.828
0.034
0.069
0.948
(50)
(50)
(50)
a: 0.194
0.880
0.990
0.060
b: 0.806
0.060
0.010
0.940
c:
0.060
(30)
(30)
(30)
a: 0.217
0.933
0.917
0.017
b: 0.783
0.067
0.083
0.983
(50)
(50)
(50)
a: 0.150
0.700
0.990
0.120
b: 0.850
0.210
0.010
0.880
c:
0.090
(29)
(29)
(29)
a: 0.155
0.914
0.966
0.086
b: 0.845
0.069
0.034
0.914
c:
0.017
(29)
Palm Canyon
(49)
Salt Creek
(30)
(50)
Simone/McCallum
Cleveland Street
drain
(29)
(Continued)
33
Table 2. (Continued)
(30)
(30)
(30)
a: 0.172
0.900
0.933
0.167
b: 0.828
0.050
0.067
0.833
C:
0.050
(35)
(35)
(35)
1.000
1.000
0.257
Shoreline pool
(29)
Living Desert
(35)
a: 0.700
0.743
b: 0.300
(39)
(39)
(39)
a: 0.410
0.885
1.000
0.154
b: 0.590
0.115
WLM
(39)
(68)
(68)
(38)
a: 0.055
0.647
1.000
0.158
b: 0.945
0.353
(55)
Howard Well
Dexter
0.842
(30)
(30)
(30)
a:
0.741
1.000
0.083
b: 1.000
0.259
Flowing Wells
JHS
0.846
(30)
(33)
a: 0.333
0.917
(94)
(96)
(95)
1.000
1.000
1.000
(51)
(51)
(51)
1.000
1.000
1.000
b: 0.667
Desert Botanical
Garden
(51)
a: 0.402
b: 0.598
(Continued)
34
Table 2. (Continued)
Boyce-Thompson
(53)
(83)
(83)
(88)
a: 0.462
0.831
1.000
0.114
0.538
0.169
b:
0.886
35
Table 3. Measures of genetic variation calculated from four
polymorphic loci in Cyprinodon m. macularius.
Values in
parentheses are standard errors.
Mean number
Number
of alleles
of loci
Population
per locus
polymorphic
Heterozygosity
Salt Creek
1.6 (0.2)
4
0.081 (0.050)
1.7 (0.3)
4
0.103 (0.045)
1.7 (0.3)
4
0.125 (0.054)
1.6 (0.2)
4
0.064 (0.029)
McCallum
1.7 (0.3)
4
0.111 (0.051)
Palm Canyon
1.7 (0.3)
4
0.101 (0.052)
Cleveland
street drain
Shoreline
pool
San Felipe
Creek
Simone/
(Continued)
36
Table 3. (Continued)
Living Desert
1.3 (0.2)
2
0.143 ( 0.093)
Howard Well
1.4 (0.2)
3
0.120 (0.069)
JHS
1.3 (0.2)
2
0.078 (0.055)
WLM
1.4 (0.2)
3
0.136 (0.071)
Dexter
1.1 (0.1)
1
0.069 (0.069)
1.1 (0.1)
1
0.064 (0.064)
1.4 (0.2)
3
0.104 (0.051)
Flowing Wells
Desert Bot.
Garden
BoyceThompson
37
Table 4. Heterogeneity chi-squares for allele frequencies in
four Cyprinodon m. macularius lineages. Lineages are
defined as a group of populations descended from a common
locality.
Lineage, locus
chi-square
Salt Creele
sAh-A
1.15
ns
Gpi-A
12.92
0.002
Gpi-B
5.59
0.018
Ldh-A
5.36
0.021
sAh-A
0.11
ns
Gpi-A
4.25
ns
Gpi-B
4.17
0.041
Ldh-A
0.05
ns
0.81
ns
San Felipe2
Canal Sanchez Taboada3
sAh-A
Terrace Springs4
sAh-A
58.83
<0.000001
Gpi-A
14.39
0.00075
Ldh-A
1.94
ns
(Continued)
38
Table 4. (Continued)
'Salt Creek lineage includes Salt Creek and Simone/McCallum
refugium.
2
San Felipe lineage includes San Felipe Creek and
Palm Canyon refugium.
2
Canal Sanchez Taboada lineage
includes Dexter NFH and Desert Botanical Garden refugia.
'Terrace Springs lineage includes WLM, Howard Well and
Flowing Wells JHS refugia.
39
Table 5. Results of hierarchical analysis of genetic
variation in Cyprinodon m. macularius following Wright's
(1978) formulation.
F-statistics and variance components
represent averages across four polymorphic loci (see Table
1)
.
Comparision
Variance
Fxy
component
X
SITE
-
LINEAGE
0.059
0.074
SITE
-
AREA
0.116
0.137
SITE
-
TOTAL
0.095
0.115
LINEAGE
-
AREA
0.057
0.068
LINEAGE
-
TOTAL
0.036
0.044
AREA
-
TOTAL
-0.021
-0.025
40
Figure Captions
Figure 1. A. Outline of historic range of desert pupfish
(Cyprinodon m. macularius; dotted line). Filled circles
represent extant localities of the Rio Sonoyta pupfish (C.
m. ssp.), and the open circle shows the single natural
locality for Quitobaquito pupfish (C. m. eremus). Modified
from Minckley (1973), Miller & Fuiman (1987), and Marsh &
Sada (1994).
B. Map of lower Colorado River and Delta with
names of localities mentioned in text.
Figure 2. History of captive C. m• macularivs populations
founded from collections on the lower Colorado River delta
and Salton Sink. The dashed line separates wild from
captive populations, and dotted lines indicate extirpated
populations. Numbers in parentheses are number of founders
for each population. Arrows indicate direction of fish
transfers. Dates near boxes and arrows indicate times of
fish transfers.
Figure 3. Proposed management plan for captive C. m.
macularius populations. Arrows indicate the direction of
gene flow. See text for details.
41
CLEVELAND STREET DRAIN
SHORELINE POOL
A SALT CREEK
ARIZONA
BAJA
CALIF9RNIA
NORTE
Oa
itifrei
SONORA
CANAL
SANCHEZ
TABOADA
1/41
L DOCTOR
r
SANTA
CLARA
SLOUGH
0
50
KM (APPROX)
GOLFO
DE CALIFORNIA
WILD SOURCES
SALTON SINK
LOWER COLORADO DELTA
CANAL
SANCHEZ
TABOADA
APR 1983
NOV 1987
WLM
(64)
1984-1985
BoyceThompson
D. Botanical
Gardens
(260)
.
3
DEC 1976
Dexter
NFH
(280)
(>110)
1984-1986
Living
Desert
(10)
i
1977
BoyceThompson
(126)
JUNE 1970
MAY 1986
(75)
DEC 1983
Deer
Valley HS
(60)
Howard
Well
(150)
SPRING 1986
Flowing
Wells JHS
(60)
SAN
FELIPE
CREEK
SALTON
SEA
TERRACE
SPRINGS
CAPTIVE
POPULATIONS
l
i
SALT
CREEK
A—UG-1981
•
Chino
(77)
SEPT 1981
Palm
Canyon
(77)
Simone/
McCallum
(78)
FALL 1987
SEPT 1981
(289)
43
INSTRUCTIONS TO AUTHORS
Submission of papers
Submission of a manuscript implies that it is not being considered contemporaneously for publication elsewhere.
Submission of a multi-authored manuscript implies the consent of all the participating authors. All papers will be
independently refereed.
Contributors from North, South and Central America should submit their papers to Dr A. P. Dobson, Department of
Ecology and Evolutionary Biology, Princeton University, Princeton, NJ 08544-1003, USA; from the Indo-Pacific region,
including Australia and New Zealand, to Dr T. R. New, Department of Zoology, La Trobe University, Bundoora, Victoria
3083, Australia; and from UK, Europe, Africa and the rest of the world to Dr B. N. K. Davis, Brook House, Easton,
Huntingdon, UK, PE18 OTU; or to a member of the Editorial Board or the publisher, for submission to the Editors.
Types of contributions
Original papers; review articles; short reports; announcements; book reviews; letters to the editor.
Manuscripts
Three copies should be provided, in double-spaced typing on pages of uniform size, with a wide margin at the left.
Generally, the size of the manuscript should not exceed 10 000 words or about 20 printed pages. Each paper should be
provided with an Abstract of about 100-150 words, reporting concisely on the purpose and results of the paper, and
with five keywords for use by Abstract services. Authors of papers in French or German are required to submit an
English translation of the Abstract. A suggestion for a running title (less than 60 characters long) may be given.
Wherever possible, authors should consult an issue of the journal for style and layout. The Editors reserve the right
to adjust style to certain standards of uniformity.
The SI system should be used for all scientific and laboratory data; if, in certain instances, it is necessary to quote
other units, these should be added in parentheses. Temperatures should be given in degrees Celsius. The unit 'billion'
(109 in America, 1012 in Europe) is ambiguous and must not be used.
Tables, references and legends to illustrations should be typed on separate sheets and placed at the end of the
paper. Footnotes should be avoided if they contain information which could equally well be included in the text.
Disks
For papers produced using a word-processor or TEX, please submit a disk with the final revised version of the manuscript.
The file on disk should correspond exactly to the hard copy. The operating system and the word-processor used should be
specified clearly.
Illustrations or chemical structures in electronic format may be supplied provided that the file format and the
program used to produce them is clearly indicated and that a hard copy is also supplied.
More detailed guidelines and further information are available from the publisher.
References
References to published work should be indicated at the appropriate place in the text, according to the Harvard system (i.e.
using author(s) name(s) and date), with a reference list, in alphabetical order, at the end of the paper. The list should
give name(s) and initial(s) of author(s), the year of publication and the exact title of the paper or book. For journals
there should follow the journal title, volume number, and initial and final page numbers of article. For books
there should follow the name(s) of the editor(s) (if appropriate), the name of the publisher and the town and year of
publication. Where appropriate, initial and final page numbers should also be quoted. All references in this list should
be indicated at some point in the text and vice versa. Unpublished data or private communications should not appear
in the reference list.
Illustrations
The original and two copies, which may be of a reduced size, of each illustration should be provided. Line drawings
may be submitted in any medium providing that the image is black and very sharp. They should preferably all require
the same degree of reduction; large diagrams, more than four times final size, are discouraged due to handling
difficulties. The type area of the journal is 177 mm wide x 250 mm deep, in two columns per page, each 85 mm wide,
and lettering should therefore be large enough to be legible after reduction of the illustration to fit (ideally 7pt lettering
after reduction). Photographs should be submitted as contrasting black-and-white prints on glossy paper. Each
illustration must be clearly numbered and the name(s) of the author(s) of the paper written on the reverse side.
Proofs
The author (or the selected author where several are involved) will receive a set of proofs for checking. No new material
may be inserted in the text at the time of proof reading unless accepted by the Editors. All joint communications must
indicate the name and full postal address of the author to whom proofs should be sent.
Page charges and offprints
There will be no page charges. Twenty-five offprints of each paper will be supplied free of charge. Additional copies
can be ordered at current printing prices.
ELSEVIER SCIENCE LIMITED
The Boulevard, Langford Lane, Kidlington, Oxford, UK, 0X5 1GB