167 The Dy amics of a Spadefoot T ad (Spea multiplicata and S. bombifrons) Hy ridization System

Marie A. Simovich

University of San Diego and The San Diego Natural History Museum

INTRODUCTION of normally efficient pre-mating isolating mecha­ nisms (Forester, 1969,1973; Frostand Platz, 1983; The phenomenon of hybridization has long Gartside, 1980 ; Martof, 1961). These studies have fascinated evolutionary biologists. Over the years examined how variation in factors such as habitat numerous questions concerning the geographical disturbance or breeding site condition can result in location, structure, and longevity of hybrid zones variation in the proportion of mixed-species matings. (D. Woodruff, 1973) as well as the causes and Extensive literature also addresses the role of post­ consequences of hybridization have been addressed mating selection, through differences in fertility, (ex. Endler, 1977, 1982; Barton, 1979; Barton and fecundity, and development on the survival of hy­ Hewitt, 1981, 1985; Hewitt, 1988; Moore, 1977). brid offspring (e.g. Brown, 1967; Forester, 1969, This work has revealed considerable variation in 1975; Frost, 1982 ; Sattler, 1978; Thornton, 1955). hybrid systems. Not only do systems differ from The interplay of pre- and post-mating selection, one another, but the interactions of a given species however, has seldom been evaluated directly in pair can vary spatially and/or temporally nature. In most cases, the dynamics of hybrid sys­ (Templeton, 1981), seemingly as a result of differ­ tems have been inferred from frequencies of geno­ ences in the responses of pure and mixed genotypes typic clas ses sampled at a single stage (usually the to the variable selectivity of spatially or temporally adult) or from breedings in the laboratory. Such patchy environments. The result is a complex or inferences may give an incomplete or even mis­ long-lived hybrid zone (e.g. Frost, 1982; Frost and leading picture, for, while revealing something Bagnera, 1977; Gartside, 1980; Gerhardt et al. about the result of the hybridization, single-stage 1980;Gollmann, 1991; Harrison and Arnold, 1983; samples tell us little or nothing about its cause. For Hillis, 1981; McDonnell et al. 1978; Mecham, this reason, information on the contribution of a 1960; Patton et al. 1979; Platz, 1981; Rand and number of aspects of selection operating through­ Harrison, 1989). These complex systems are par­ out the life cycle is important to the evaluation of ticularly informative because the genetic make-up hybridization systems. This paper is one part of a of each component species remains comparatively study of a complex hybrid zone between two spe­ uniform. This consistency makes it easier to evalu­ cies of spadefoot toads in Arizona. In it, pre- and ate the role of external forces (e.g., crowding, post-mating aspects of selection, including assorta­ habitat variation) acting both before and after mat­ tive mating, differences in fertility and fecundity, ing, and their effect on the outcome of hybridiza­ and differences in survival and developmental rate, tion. are investigated. Among anurans, numerous studies have investigated factors responsible for the breakdown la. C1 Y Herpetology of the Norl:h America~ Deserts~p p. 167-1 82 i n P. R. Brow n & S.W . Wr ig ht (eds.), Southwestern Herpetologl sts Socl et y . S 'peci al publ i cat i on #5. 168 MARIE A. SIMOVICH The System: tor/scavengers with enlarged beaks and hypertro­ The ranges of the spadefoot toads Spea phied jaw muscles. multiplicata and S. bombifrons overlap widely In the San Simon Valley, the ponds may h.~. (Brown, 1976; Stebbins, 1985). Although consid­ from three days to over two months, depending on ered to be the most distantly related species within their size and the amount of rainfall. Inyears of low the genus (see for discussion Wiens and Titus, or patchy rainfall, however, ponds frequently dry 199 1), the species are known to hybridize in a before all or sometimes any tadpoles metamorphose num ber of areas to varying extents (Brown, 1976; (Pomeroy, 1981, personal observation). Also, the Forester, 1973; Hughes, 1965; Sattler, 1978, 1985), tadpoles are exposed to a number of predators, resulting in a complex hybrid system. One OJ these another source of high mortality, and there is pre­ areas, the San Simon Valley of southeastern Ari­ sumably significant selection pressure for rapid de­ zona, is part of a suture zone (sensu Remmington, velopment (Caldwell et al., 1980; Creuser and 1968), an area where biomes intergrade and conge­ Whitford, 1976;Crump, 1989; Licht, 1974; Mayhew, ners meet and frequently hybridize. The valley itself 1965; Newman, 1988a, 1988b ;Sarnlitsch and Wilbur, has a large number of ponds and cattle tanks where 1988 ;Travis, 1980, 1983; Wilbur, 1977; Wilbur and the toads breed. Collins, 1973; Woodward, 1987). Any differences Breeding in ephemeral desert ponds con­ in development between pure or mixed genotypes strains the spadefoot's life cycle into precise should be exposed to immediate selection. synchrony with a very temporary aquatic environ­ Several aspects of the hybridization of these ment. The toads breed "explosively" (Wells, 1977), species facilitate the study of this system. First, the emerging from their burrows and going to tempo­ pure species, hybrids, and backcrossed individuals rary ponds and playas that fill on the first night of are easily and accurately identifiableelectrophoreti­ heavy summer rains (Bragg, 1965; Dimmitt and cally (Simovich, 1985). Second, the toads breed Ruibal, 1980; Forester, 1969; Ruibal et al. 1969). essentially on one night, allowing adults to be id The breeding congresses that form are often dense tified. Third, cohorts of tadpoles can be sampled and contain both species (some contain other genera repeatedly to metamorphosis, and comparisons be­ oftoads as well) (Bragg, 1965; Brown, 1976; Creuser tween stages can be made to assess the success of the and Whitford, 1976), taking part in what has been various pure and mixed types with respect to several termed "scramble competition" for mates (Wells, factors of selection. 1977). The females respond positivel y to the calls of conspecific males, and the males of the two species Background: call from different positions in the pond (Forester, Other investigations of this system (see 1969, 1975). Males are quite indiscriminate and will Simovich, 1985; Simovich and Sassaman, 1986; mate with either species or hybrids (Blair, 1958; Simovich, et al., 1991) provided the background Bragg, 1965; and personal observation). Under data for the present study. First, protocols were theseconditions, the effectiveness ofcall and calling established to identify genetic classes (pure species, position as mechanisms of assortative mating may hybrids, etc.) by means ofallozymes (Simovich and be much reduced (Wasserman, 1957; Wells, 1977), Sassaman, 1986). The distribution of geneticclasses and females may be grabbed by an incorrect male en in populations of tadpoles in 27 ponds in the valley route to a conspecific. was then determined over a three-year period. The The eggs are laid during the breeding con­ frequency of genetic classes was found to vary both gress and hatch the next day. The tadpoles develop from pond to pond and within a pond from cohort to and metamorphose in just three to four weeks. The cohort. While pure S. multiplicata was generally the tadpoles of these two species are dimorphic, occur­ most common class, in some ponds the two pure ring in an omnivore or a carnivore morph (see species were equally abundant. The frequenc­ Pomeroy, 1981). Omnivores are scavenger/grazers the hybrid classes varied from 0 to 40% (Simovicn, of normal morphology, while carnivores are preda­ 1985).

. Southwestern Herpetologists Society SPADEFOOT TOAD HYBRIDIZAnON 169 The role of pre-mating isolation was also typic class at numerous points in the cycle. addressed. Breeding adults were sampled in several breeding congresses. Assortative mating was effi­ METHODS cient in large or uncrowded ponds; there were few mismatches, and few hybrid tadpoles resulted. Over three years, toads in three ponds were Crowded breeding conditions (due to low rainfall or sampled periodically to metamorphosis. Six more the use of small ponds) decreased the efficiency of cohorts at two other ponds were sampled for at least assortative mating, increased mismatching, and in­ some consecutive post-breeding stages. Ponds are creased the generation of hybrids and backcrossed coded and locations are given in Simovich (1985). offspring. Most of the backcrosses resulted from a Figure 1, a schematic outline ofthe anuran life cycle, maleS. multiplicata (themorecommon ofthe paren­ indicates the stages sampled and used in predicting tal species) mating with a female hybrid. Although values, as well as the relevant stage comparisons the frequency of the mixed-genotype classes in made to determine differential performance of ge­ tadpoles was correlated with the frequency of mis­ notypic classes for the several points in the life matches by breeding adults, the proportion ofhybrid cycle. tadpole types could not be predicted precisely solely These classes are defined and abbreviated as from the degree of adult mispairing. The discrep­ pure S. multiplicata (M), pure S. bombifrons (B), ancy suggested that there was selection after as well hybrid (H referring to F! hybrids only) backcross to as before mating (Simovich, 1985). S. multiplicatus (BKM), backcross to S. bombifrons Analysis of intraspecific and interspecific (BKB), and double backcrosses (DBK), the off­ crosses from this area (Sirnovich et al., 1991) showed spring of two or more backcrosses. "Mixed" denotes that hybrid males are sterile and that hybrid females, hybrids plus all backcross classes. All animals were although fertile in backcrosses, produce only 45% as identified to genotype by electrophoresis of tissue many eggs as the pure species produce. This differ­ samples, which were scored for four allozyme sys­ ence in reproductive potential could affect the fre­ tems representing four unlinked diagnostic gene loci quency of tadpole classes by limiting the production (Simovich and Sassaman, 1986). of F hybrids. Mixed genotypes, however, are com­ 2 mon in the zone of hybridization. Frequencies of Sampling: hybrid adults range from 0 to 31 % and of hybrid Ponds were monitored nightly until they first tadpoles from 0 to 40%. filled with water. After filling, toads were allowed Laboratory experiments further showed that to enter undisturbed. Late in the evening, amplexed on a diet including live food (fairy shrimp), as in pairs were sampled by hand net. It is estimated that natural ponds, hybrid tadpoles survive in higher the samples included 75-80% of the pairs in most numbers, develop faster, and enjoy better net suc­ cases. Tissue samples were taken (toe clippings), cess to metamorphosis than do tadpoles of either and the pairs were released to continue breeding. pure species. In the lab, hybrid toad lets were inter­ The genetic class ofeach partner was determined by mediate in size between toadlets of the parental allozyme analysis, and the frequencies of all combi­ species of the same age (Simovich et al., 1991). nations were calculated. Ponds were checked for Such differences should give rise to differences in several succeeding nights and aftersubsequentstorms the success of the various genotypic classes and to determine if there were further breeding con­ could allow substantial numbers of hybrids to per­ gresses. Although a small number of males may call sist. for a few nights after a rain, observations at these and In the present study the possibility that both numerous other ponds in the area revealed only one pre- and post-mating aspects ofselection were oper­ instance ofsignificant mating after the first congress ating was investigated in the field by following (>0.5%). Furthermore, subsequent rains did not cohorts of larvae to metamorphosis, and assessing bring breeding congresses back to a pond once a full the relative performance or success of each geno­ congress had convened at that pond.

Herpetology of the North American Deserts 170 MARIE A. SIMOVICH 7 , , / "' ...... " "" ...... I " " --..... " <, ,," , 6...... " <, / I...... ' " I...... " " ...... --..... " I 3 ...... , / " I""",/" " ...... --...... ,," I'" <; <, ,," I"'/" '-,-'\. <, ADU LTS .... EARLY... LATE R EARLIE ST TADPOLE S TADPOLE S TOADLET S ------5 I I I I 1 I I 11 a I 1 1 I 1 1 14 1 I I UN MODIFIED 1b I I I I PREDI CTED SLO W LATER TAD POLE S MODIFIED DEVELOPER S ... TOA DLET S

I PR EDICT ED 1 TAD POLE S 2

Figure 1. Summary of the Spea life cycle, indicating the Ponds were then checked daily to determine stages sampled and com par isons made. when tadpoles began to metamorphose. Toadlets 1) A comparison of sampled and predicted early tadpole class metamorphosing early (emerging on the first I distributions a) Sampled versus unmodified predictions , gener­ days of metamorphosis) were collected from under ated from adult pair combinations, incorporating assortative rocks and debris at the periphery of the pond. These mating only. were then typed as well. Consecutive samples of b) Sampled versus modified predictions , incorporat­ metamorphosing individuals were taken at five-day ing both assortative mating and laboratory measurements of intervals from some ponds. Not all ponds could be differences in reproductiv e potential. 2) Difference in goodness-of-fit between la and lb. sampled sequentially through tadpole metamor­ 3) A comparison of early versus late tadpole class distribu­ phosis because of population size. In some ponds tions. only the group emerging first was sampled. The 4) A comparison of earliest metamorphosing toadlets and criterion used to determine if toadlets had metamor­ slow developers. . phosed recently (24 to 48 hrs. from emergence) was 5) Time sequence of emergence of classes as toadlets. 6) A comparison of early tadpoles versus earliest metamor­ the possession of a tail stub still long enough to phosing toadlets . wiggle. Toadlets were measured for size (snout­ 7) A comparison of adults versus early metamorphosing vent length) at the completion of metamorphosis, toadlets . and the mean size of genetic classes was compared. For this, only toadlets that had lost their tadpole Following breeding, early tadpoles (7 to 10 beak and much of the tail (and were therefore of days old) were sampled from ponds by hand net and toadlet size and form) were used (toadlets' snout­ typed electrophoretically. Late tadpoles (with hind vent length decreases while they lose their beaks legs well developed on at least some individuals and tails). Developing tadpoles remaining in the and usually 12 to 20 days old) were later sampled pond after the first group had metamorphosed were and typed in the same manner. The exact day of then sampled and typed in the same manner sampling was determined by the rate of develop­ earlier tadpoles. For lack of a better term these are ment of tadpoles in the various ponds. called slow developers.

Southwestern Herpetologists Society SPADEFOOT TOAD ffiTflRIDIZATION 171 between predicted and sampled distributions, the F Analysis: ratio (ratio ofvariances) of the Chi Square values of From the matrix of mated pairs, the ex­ the two prediction comparisons was determined pected arrays of offspring classes were generated (Fig. 1, comparison 2). for each cross (given the Mendelian assortment of The differences in post-em bryonic survival the unlinked loci). Each mating was then assigned through the tadpole stage were then evaluated by two offspring, which were allocated proportion­ comparing the frequencies of the classes in the ately into the classes expected from that cross. This samples of early and late tadpoles (Fig. 1, compari­ assignment oftwo replacements maintains the same son 3). The extent of differences between classes in sample size as in adults. The predicted distribution developmental synchrony was next assessed by was based initially on the assumption of equal comparing the arrays for the early metamorphosing fertility, fecundity, and viability of classes and all tadpoles to those for the slow developers (Fig. 1, crosses. It does, however, account for the accuracy comparison 4). This is not a direct measure of ofidentification ofbackcrosses arising from the use fitness, since concurrent rather than sequential of only four markers (87 %) (Simovich and samples were compared. Where available, samples Sassaman, 1986) and the degree of assortative of tadpoles metamorphosing at consecutive five­ mating observed. day intervals were also compared to determine the For each matrix, the combined effects of time sequence of the metamorphosis of each class differential fertility, fecundity, and early aquatic (Fig. 1, comparison 5). For a further indication of viability were then evaluated as an early composite differences in developmental rate, early metamor­ component of selection by comparing (Contin­ phosing toadlets and early tadpoles were compared gency Chi Square) the observed distributions of (Fig. 1, comparison 6). A final measure of differ­ early tadpole types with the predicted values gener­ ence in the representation of the genotypic classes ated as described above (Fig. 1, comparison 1a). was assessed by comparing the distribution ofclasses Next, to account for differences in the fertil­ among early metamorphosing tadpoles to that of ity and fecundity of adult hybrids as determined the breeding adults (Fig. 1, comparison 7). No under laboratory conditions (Simovich, et al. 1991), comparison of adults versus all metamorphosing a modified predicted distribution of classes was toadlets was made because of the prolonged time generated and compared (Contingency Chi Square) members of some classes took to develop and the with the sampled distribution (Fig. I, comparison damage this would have on the populations. Some 1b). For this modification, two correction factors samples of tadpoles from these and other ponds were employed. First, to account for the sterility of were also scored for carnivore/omnivore morphol­ hybrid males, predicted distributions eliminated ogy. the offspring contribution of that proportion of For each comparison, heterogeneity among pairs involving F, hybrid males. Second, to account the full arrays of classes at various stages of devel­ for the reduced fecundity of hybrid females, the opment was tested by Contingency Chi Square predictions were modified by multiplying the off­ analysis (Chi Square values were considered sig­ spring contribution of pairs with hybrid females by nificant at p<.05). Also, for each comparison, a 0.45 (value from Simovich et al., 1991). These two relative performance value (or difference value in modifications had little effect on the predicted the case of developmental rate) was calculated for proportions of hybrids (since most were produced each class relative to the population: by direct F j crosses of the pure species) but they W =sampled frequency / predicted frequency drastically reduced or eliminated expected propor­ or, tions of some multiply mixed classes. The modifi­ W =frequency in a stage/ frequency in preced­ cations increased the expected proportions of the ing stage. pure species. To determine if the modifications The value W is a measure of the change in relative produced a significant decrease in the difference frequency ofa class compared to that in a preceding

Herpetology of the North American Deserts 172 MARIE A. SIMOVICH stage or of a difference relative to a predicted value. for heterogeneity in the performance relative to S. Performance values greater than 1.0 indicate a multiplicata of each class at each developme higher than predicted frequency of a class or in­ stage are given in Appendix 1. creased relative representation of a class through the cycle. RESULTS Comparisons were then standardized by comparing the performance ofeach class relative to Comparison of sampled versus predicted that of the predominant species, S. multiplicata. arrays shows significant differences among classes Significant heterogeneity in performance relative which includes fertility, fecundity and hatchability. to S. multiplicata in individual ponds was deter­ The raw proportions of the classes of breeding mined by Contingency Chi Square analysis. To adults alone are not good indicators of those of the then evaluate the net performance relative to S. early tadpoles. Predicted proportions of tadpole multiplicata over all tests, a Z statistic was calcu­ classes based on mated pair combinations alone lated as Z =Chi square/N"(Everitt, 1977). For this differ significantly from observed proportions (Con­ statistic, Chi values for each separate stage-stage tingency Chi Square) (Table 1). Modifications in­ comparison relative to S. multiplicata were summed corporating laboratory measurements of reproduc­ over all N cohorts. Values greater than 1.96 indi­ tive differences between adult classes improve the cate net performance better than that of S. fit (Table 1). The improvement is due primarily to multiplicata over all cohorts for that comparison. a reduction in the total number of genetic classes Values of -1.96 or less indicate poorer net perfor­ expected. This reduction results from the elimina­ mance. A summary of the results of individual tests tion of crosses involving hybrid males and a reduc­

I Table 1. Summary of comparisons of predicted and sampledtadpole class distributions. Contingency Chi square test for heterogeneity between: 1). sampled versus predicted distributions which incorporate assortative mating, 2). sampled versusmodified predicted distributionswhichincorporate assortativemating and differential reproduction, 3). F-ratio between first and second test to determinesignificance of decrease in heterogeneity, an asterisk indicates significance at 0.05 level.

NUMBER OF TYPE CLASSES CHI SQUARE F

MODIFIED POND PREDICTED PREDICTED SAMPLED 1 2 1/2

M80 6 4 3 90.27 45.88 1.97

M81 6 4 5 130.20 16.20 8.04*

M82(l) 6 5 3 113.14 28.11 4.02

M82(2) 6 4 4 164.90 17.06 9.67*

G80 6 4 4 48.03 12.91 3.72

G82 6 3 1 31.28 9.78 3.20

S81 5 5 4 19.46 18.54 1.05

Southwestern Herpetologists Society SPADEFOOT TOAD HYBRIDIZATION 173 tion in the expected frequency of offspring from crosses involving hybrid females. These classes of offspring (e.g., double backcrosses) are, in fact, rare or absent in the samples. Incorporation of fertility and fecundity data thus brings the predicted num ber 1.0 of classes closer to the observed number in most cases. In one congress, no hybrid males were found in pairs; therefore, the modification does not change .8 (/') the prediction substantially. Genetic types result­ :=l ing from backcrosses to S. multiplicata are not -C'Cl generally eliminated but are reduced. These back­ o .6 crossed classes are the major classes affected by the Q. low fecundity of hybrid females in crosses to S. :=l multiplicata males. The expected frequencies of E .4 hybrids and S. bombifrons are not changed as radi­ Cf) cally by modifications, since these tadpoles are produced by the rarer pure species and F crosses .2 j >­ that, in the laboratory, do not differ in reproductive () potential. Z w .0 Among the early larvae S. multiplicata is :::J over-represented relative to expectations (Fig. 2). a .4 w Over all comparisons (Z values, summed Chi a: Square), S. bombifrons, hybrids, and backcrosses to u... -0 S. multiplicatus are significantly scarcer (Fig. 3-1b) w a: .2 > CD than expected. The deficiencies of tadpoles of S. fo- >­ bombifrons cannot be explained adequately on the -c ::r: -l basis of laboratory measurements of fecundity be­ W .0 a: cause the two pure species do not differ in fecundity .6 (under laboratory conditions). Thus, some other factor seems to be affecting S. bombifrons offspring (/') c: adversely in a way not detectable in the laboratory. o .4 ~ In four of five cases they performed significantly '+­ .0 more poorly than did S. multiplicata (Appendix I­ E Ib). Laboratory data are also inadequate to account o .2 .0 for all adverse selection affecting hybrids. But, hybrids' reduced reproductive potential can explain some of the difference between the values for the .0 backcrossed classes. There is a slight difference in survival be­ A P PM E tween classes when early and later tadpoles are compared. Z values (summed Chi Square) show Figure 2. Relative frequencies of classes of adults (A), that overall S. bombifrons performs significantly predicted tadpole arrays based on mated pairs (P), modi­ better than S. multiplicata. In no case did S. fied predictions incorporating reproductive differences (PM), and sampled early tadpoles (E). The representa­ multiplicata tadpoles show significantly better sur­ tion of backcross classes is not shown here. Each line vival (increased representation) than did S. represents the sample from a single pond. bombifrons tadpoles. In the laboratory post-embry­ onic viability of the offspring of backcrosses to S.

Herpetology of the North American Deserts 174 MARIE A. SIMOVICH multiplicata is also somewhat lower under some conditions (Simovich, 1985). For hybrids ther­ H B BKrv1 5 a .fair degree of heterogeneity among cohorts 111 survival, but the Chi Square values are insignificant o relative to S. multiplicata (Appendix 1-3) and show - 5 no net difference. Backcross survival is somewhat • • worse (Fig 3-3). - 1 0 When the earliest toadlets are compared to EARLY/MODIFIED PREDICTED slow developers, S. bombifrons, hybrid, and back­ TADPOLES cross-M classes are overrepresented relative to S. multiplicata over all comparisons (summed Chi 1 0 Square, Fig. 3-4). The relative frequency of S. 5 • bombifrons and hybrids in samples of the earliest toadlets is usually higher than in samples of the o slow developers. Inon!y one case was S. multiplicata - 5 significantly better (Appendix 1-4). In one pond, S82, the tadpoles metamorphosed quickly and syn­ LATE/EARLY TADPOLES chronously in less than two weeks. Slow develop­ ers were too sparse to sample; however, the fre­ 1 5 • quencies of hybrid and S.bombifrons toadlets in the 1 0 • first sample were higher than that ofS. multiplicata and showed an increase relative to the early tadpole 5 samples. 0 In those ponds from which consecuL samples are available (Fig . 4), hybrids and S. - 5 bombifrons are in higher than initial frequency in EARLY METNSLOW the samples metamorphosing earliest, and the later DEVELOPERS samples are composed mostly of S. multiplicata, indicating that hybrids and S. bombifrons complete 20 metamorphosis sooner and that theirsurvi vors leave • the aquatic environment sooner. 1 5 • In the comparisons of the proportions of 1 0 classes from early tadpoles to early metamorphosis, hybrids were more successful than S. multiplicata 5 in six of ten cases (five significant), than S. 0 bombifrons in six of seven cases (all significant), and than backcrossed to S. multiplicata in two of - 5 eight cases (both significant) (Appendix 1-6). In EARLY METNEARLY only one case was the poorer representation of S. TADPOLES bombifrons relative to S. multiplicata significant.

Figure 3. Summary of performance differences relative to S. • • • muttiplicata. Plotted are Z statistic values which summarize ~ j performance comparisons over all N cohorts. Where Z is greater than 1.96, there isoverall better performanceofa clar - 5 relati ve to S. multiplicata; where Z is less than -1.96 there . EARLY METNADULTS overall poorer performance. B = S. bombifrons, H = hybrid, BKM =backcross to S. multiplicata.

Southwestern Herpetologists Society SPADEFOOT TOAD HYBRIDIZATION 175

1.0 S.multiplicatus HYBRID S. bombifrons

>­ .8 0z w ::> a .6 w 0: LL W > .4 l­ e:( -J W 0: .2

.0

E M1 M3 E M1 M3 E M1 M3 STAGE

Figure 4. Relative frequency changes between early tadpoles and the earliest and subsequent metamorphosis samples. Size At Emergence: Shown are the frequencies of classes in early tadpoles (E), and S. bombifrons toadlets were larger than S. for all metamorphosis samples available. When sequential multiplicata toadlets in all ponds containing both metamorphosis samples are available, these are referred to as species. Sizes of hybrid toadlets were generally MI, M2, and M3. MI is the earliest metamorphosing toadlet sample and is available for the most ponds . intermediate between those of the pure species of the same age but overlapped with that of S. bombifrons in some ponds (Fig. 5). This is the same S. multiplicat« was frequently under-represented, rank order found in laboratory experiments and hybrids and S. bombifrons were over-repre­ (Simovich, 1985). In only one pond did hybrids sented among the early metamorphosing toadlets in overlap with S. multiplicata. The number of hy­ comparison to their initial frequency among the brids in that sample (Shot 82-2) was quite low, and early tadpoles (Fig. 4). Summed Chi square values hybrids were larger than S. multiplicata in the first (Z) indicate that over all ponds sampled, S. sample of the sequence. Backcrosses to S. bombifrons and hybrids performed better than S. multiplicata, when present, may also be larger than multiplicata, while backcrosses to S. multiplicata S. multiplicata. The size of toadlets within a given did not differ (Fig. 3-6). The results indicate that S. class varies from pond to pond. But, when sequen­ bombifrons and hybrids generally develope faster tial samples are available within a pond, the late than does S. multiplicata. emerging S. multiplicata (the species for which we Comparison between the breeding adults have the most data) are not larger than the early and early metamorphosing toadlets show no clear emergers of the same species, and thus are not differences in the repres entation of classes. benefiting from a prolonged aquatic period by meta­ morphosing at a larger size.

Herpetology or the North American Deserts .... -...I Figure 5. Variation in toadlet size between classes and between ponds. Average snout-vent length (mm.) and 95% C.L for emerging toadlets. Asterisk indicatescases 0\ wheremultiplerange testsofANOVA indicate asignificant differencedespiteoverlapof the95%C.L. Dotted linesconnectlike genotypesappearinginsequentialsamples from individual ponds of the same cohort. Abbreviations below are individual pond codes. Not all classes were present in all ponds.

28 1', ------­ f

26 B 28 ,i ------­ , l B H rJJ B o ....c 26 B e­ :2 24 B ~ t + OBK ~ ~ H :2 35 89 til.... t > ~ 8 ., I + B ~ 24 '=' = f­ H BKM S3 ::t (9 22 34 H ~ ~ B T / ., I / z / f- / rJJ w / '[ 102• M CJ 2 2 8 / ,/" .-J H ~ o 30:t z M M 0' f­ + w o 52 .-J H.+/ H t Q3. Z 20 23 +, M J+ ;:5 til 8 23 w 37+ f­ +,,69 (J rz Z 20 9 111\ , , J/50 rJJ > M " , J 4 BKM ::r:: I w M 36 ", , J g f­ > , I I ;" I .... 18 4 I­ '< 6 M 117• T' -. M T M z 6 18 Cf) z 18 " " " ' " M! 1 t----~-+ (f) 62• M 4 16 133.-- -r 16 55 16 13 40+ M 4 4 14 14 META2 META1 M3 135• META1 M3 META2

ALB ALB SHOT82 ALB82 GAZP82 MHL82(2) 80 81 SP ADEFOOT TOAD HYBRIDIZAnON 177 Carnivore Morphology: breeding. This pre-mating aspect of selection was In all mixed species ponds analyzed, the addressed in this study by sampling the actual proportion of S. bombifrons and hybrid tadpoles mated pairs from a particular pond and then using with the carnivore morphology was higher than that this information to predict the frequency of geno­ in S. multiplicata. Spea multiplicata never develops typic classes among the resulting tadpoles. Be­ carnivores in proportions as high as S. bombifrons cause the observed frequencies of classes differ does, even in those ponds where it is the only Spea significantly from these predictions, post-mating present (Fig. 6). However, carnivores are some­ selection must be involved. what more frequent in S. multiplicata itself when A significant portion of the discrepancy that species faroutnumbersS. bombifrons (e.g., Alb between observation and prediction can be ac­ 81) or when no S. bombifrons is present (ponds of counted for by post-mating selection in the form of pure S. multiplicata). Carnivore tadpoles from reduced reproductive potential of hybrids as mea­ backcrosses of hybrids to S. multiplicata are more sured in the laboratory (Simovich et al., 1991). frequent than in S. multiplicata (more as in S. Modifying the predictions to reflect the difference bombifrons). in hybrids' reproductive potential qualitatively ac­ counts for the low frequencies of some backcrosses DISCUSSION and suggests that the fertility differences measured in the laboratory are real and likely to be important Ithas been shown that much ofthe variation in nature. There is, however, a further portion of the in the level of hybridization and in the frequencies discrepancy that this investigation has notexplained, of the various classes of tadpoles in this area can be the low initial representation of S. bombifrons tad­ accounted for by variation in assortative mating poles. It is possible that the early success of S. (Simovich, 1985). Mating is more random and bombifrons in nature is more closely tied to envi­ more tadpoles of mixed genotype are produced ronmentalconditions than is thatofthe otherclasses. when ponds are small and crowded at the time of Despite adverse selection in early compo­

1.0 ~ B 0..•. BKB W OH [( .8 BKM 0 0 > EJ OSK Z o; OM « .6 (.,! o z 0 f­ .4 [( 0 o, 0 ':. ' o; .2 .,. a, ..

PURE MPONDS RIGGS82 SEA81 ALB80 SHOT81 ALB81 216 18 74 203 98 56 198 160 142

Figure 6. The proportion within each class (in field samples) with carnivore morphology. Some ponds are pure S. multiplicata, others are variously mixed. For mixed-species ponds, the overall frequency of S. muliiplicata classes is greater than S. bombifrons or hybrids except for Sea 81 where the relative frequencies ofS. multiplicata and S. bombifrons are about equal. For Alb 81, the clear bars indicate higher proportions in a second, later sample of tadpoles. M, B, H, BKM, as well as backcrosses to S. bombifrons (BKB) and multiple backcrosses (DBK) are indicated. Bars below zero indicate the class was present in the pond but no carnivores were found.

Herpetology of the North American Deserts 178 MARIE A. SJMOVICH nents of fitness , S. bombifrons and hybrids often does not seem to gain a size advantage by prolong­ constitute the majority of the first tadpoles to meta­ ing the larval period. morphose. Laboratory experiments had predicted Interestingly, the classes that are largest aIIU fast development of S. bombifrons but had demon­ develop fastest, hybrids and S. bombifrons, also strated that hybrids develop faster than either pa­ express the carnivore morph most frequently in the rental species when live food (fairy shrimp from field. The expression of carnivore morphology is natural ponds) is available (Simovich, 1991). The quite plastic and has been tied to a diet oflive food, variability in the expression of developmental dif­ and since carnivores tend to develop faster and be ferences in the field again raises the possibility that larger than omnivores (Pomeroy, 1981), there may the expression of these differences may vary in well be a direct relationship between the develop­ response to environmental conditions. ment of carnivore morphs and the other develop­ Several studies have shown that anuran mental differences between classes. As carnivore developmental rates can be affected by variations in morphs do not develop well in the laboratory, we do environmental conditions; these include tempera­ not currently understand the relative contributions ture, drying, food availability, and crowding (e.g., of genetics, diet, or other variables to this polymor­ see Collins, 1979; Crump, 1989; Harkey and phism. Semlitsch, 1988; Newman, 1988a, 1988b;Semlitseh Substantial numbers of hybrids can be pro­ and Wilbur, 1988; Travis, 1983, 1984; Wilbur, duced and survive under some conditions. Al­ 1977; and Wilbur and Collins, 1973). In the field though selected against early in the life cycle and in portion of this study variation in environmental subsequent reproductive potential, hybrids and S. conditions from pond to pond could not be con­ bombifrons may be favored during the larval phase trolled. The ponds all retained water throughout the because of their rapid rates of growth and develop­ larval period of the toads, though all experienced a ment. Because of this their net fitness may }-<> reduction in volume. Trends in developmental higher than that of the more prevalent speci.. rates observed in different ponds and in the labora­ particularly if ponds dry early. Furthermore, this tory were largely consistent. This strongly suggests allows substantial introgression of alleles between that there is an important genetic (phylogenetic) species at least in some areas. Hybrid females can component to developmental rate difference. The be frequent enough to produce significant numbers fitness advantage conferred by fast development to of backcross offspring (especially with the more avoid desiccation and predation (Caldwell et al., common S. multiplicata males) despite their low 1980; Collins, 1979; Licht, 1974; Mayhew, 1965; fecundity. Introgression may then constitute a Travis, 1980, 1983) cannot be ignored. In drought ready source of new variation and may have evolu­ years, this advantage could well be enjoyed by S. tionary importance within the system. bombifrons and hybrids over S. multiplicata. The intensity and direction of selection in Faster development of S. bombifrons rela­ this system can vary at different parts of the life tive to S. multiplicata is accompanied by largersize. cycle and under different conditions. The distribu­ Hybrids, in the field as in the laboratory, are inter­ tions and net fitness of each genotypic class varies mediate in size. Being larger at metamorphosis in response to the environment. The condition-­ may increase first-year survival and the probability small pond size-increasing the production of hy­ of early maturation (Collins, 1975, 1979; Martof, brids, by decreasing the efficiency of isolation 1956; Mayhew, 1965; Wilburetal.1978). In many (crowding), could also favor their survival. anurans, slower developers attain a larger size The vegetation and runoff patterns in the (Crump, 1989; Collins, 1979; Newman, 1988; and San Simon Valley have changed significantly in the Travis 1980, 1983, 1984). In S. multiplicata and S. past 100 years. The valley was once primarilv bombifrons however, this tradeoff is not apparent grassland, but now mesquite has invaded from t between or within the species. The species with the hillsides far into the valley. Rainfall in the summer longer period of metamorphosis, S. multiplicata, months of the monsoon season can be very spotty,

Southwestern Herpetologists Society SPADEFOOT TOAD HYBRIDIZATION 179 and the runoff is now diverted into a number of BLAIR, W. F. 1958. Mating call and the speciation of anuran cattle tanks installed by ranchers. A large cienaga amphibians. Arner. Naturl . 92:27-51. at the northern end of the valley has been reduced. BRAGG, A. N. 1965. Gnomes of the Night. Univ. of Penn. The result is that the toads now breed in many small, Press. Philadelphia. scattered ponds. It has been speculated that the effect of habitat modifications on hybridization BROWN, H. A. 1976. The status of California and Arizona between species can be exacerbated by drought populations of the western spadefoot toad (genus Scaphiopus). L. A. Co. MuseumofNat. Hist.Cont, inSci. (Cousineau and Rogers, 1991). These changes 286:1-15. . have altered the breeding sites, which were parti­ tioned by the two species prior to the establishment BROWN, H. A. 1967. Embryonic temperature tolerance and of the cattle industry, and may have increased the genetic comparability in two allopalric populations of the probability of pool crowding and thus mismatch­ spadefoot toad,Scaphiopus hammondi. Evolution 21:742­ 761. ing. The same conditions could increase the prob­ ability of early drying, selecting against the slower CALDWELL, J. P., J. H. THORP and T. O. JERVEY. 1980. developing species at times. Whatever the cause of Predator-prey relationships amoung larval dragonflies, the hybridization, environmental variation now, at salamanders and frogs. Occologia 46:285-289. times, facilitates the production and survival of COLLINS, J. P. 1975. A comparative study of the life history substantial numbers of hybrids and their offspring. strategies in a communtiy of frogs. Ph.D. Thesis, Univ, Although hybridization may have occurred natu­ Michigan. rally on a smaller scale before human intervention, it is quite probable that the current high levels of COLLINS, J. P. 1979. Interpopulational variation in the body hybridization in some portions of the zone are size at metamorphosis and timing of metamorphosis in facilitated by modification of water flow. the bullfrog, Rana catesbeiana. Ecology 60:738-749. COUSINEAU, M. and K. ROGERS. 1991. Observations on ACKNOWLEDGMENTS sympatric Rana pipiens, R. blairi and their hybrids in eastern Colorado. J. Herpetology 25:114-116. I would like to thank R. Tinsley and his wonderful co-workers, G. Bell, A. Chovnick and CREUSER. F. M. and W. G. WHITFORD. 1976. Ecological relationships in a desertanuran community. Herpetologica the numerous researchers at the SWRS for their 32:7-18. invaluable help catching toads in the pouring rain CRUMP, M .L. 1989. Effect ofhabitatdrying on development and counting thousands of eggs. I would also like time and size at metamorphosis in Hyla pseudopuma. to thank C. Sassaman, M. Wells, L. Baird, and P. Copeia 1989:794-797. Unitt for reviewing portions of the manuscript, B. DIMMITT, M. A. and R. RUIBAL. 1980. Environmental Coleman for typing it and 1.Ebnerand B. Henscheid correlates of emergence in spadefoot toads (Scaphiopus), for artistic contributions. This work was funded in J. Herpetology 14:21-29. part by grants from the National Science Founda­ tion, theTheodore RooseveltMemorial Fund, Sigma ENDLER, J. A. 1977. Geographic Variation, Speciation and Xi, and the Chancellor's Patent Fund (UCR). Clines. Princeton Univ , Press. Princeton N.J. EVERITT, B. S. 1977. The Analysis Of Contingency Tables. LITERATURE CITED John Wiley and Sons Inc., N. Y.

BARTON, N. H. 1979. The dynamics of hybrid zones . FORESTER. D. C. 1969. Reproductive isolation and hybrid­ Heredity 43:341-359. ization between the spadefoot toads Scaphiopus bombifrons and Scaphiopus hammondi in West Texas. BARTON, N. H. and G. M. HEWITT. 1981. Hybrid zones M.S. Thesis. Texas Tech. University, Lubbock, Texas . and speciation, p.109-145. In: W. R. Atchely and D. Woodruff (eds.). Evolution and Speciation. Cambridge FORESTER. D. C. 1973. Mating call as a reproductive University press . New York. isolating mechanism between Scaphiopus bombifrons and Scaphiopus hammondi. Capeia 1973 :60-67.

Herpetology of the North American Deserts 180 MARIE A. SIMOVICH

Appendix 1. Analysis of the six classes relative to S.multiplicata (M). For each componentthe stages or predicted values compared are indicted. For each comparison, the changes in frequency ofa class between stages or differences between expected and observed frequencies (performance value) are evaluated relative to those of S. multiplicata by Contingency Chi Square Analysis. Shown are the number of comparisons where a class increased T, or decreased J" relative to M (denominator) and the number of those differences which were significant (numerator) and the number which did not differ (=).

Component Direction Classes and Relative Comparison toM H B BKM BKB DBK

la Early r 0/2 0 0 0 0 vs J, 2/5 3/7 6/7 5/7 5/6 Mod. Pred. = Ib Early r 0/3 0 0 011 0 vs J, 2/4 4/5 4/7 1/2 011 Mod. Pred. = 3 Late r 1/6 3/6 011 0/2 011 vs J, 0/6 0/2 1/9 011 0 Early = 4 Meta r 4/4 4/5 1/2 2/2 1/1 vs J, 0/2 III 0/4 III 0 Slow = 1 6 Meta r 5/6 6/6 2/2 0 III vs J, 0/3 1/1 0/4 011 0 Early = 1 2 7 Meta r 2/2 1/3 III III 0 vs i 2/3 2/2 0/1 III 0 Adults = 1

FORESTER, D. C. 1975. Laboratory evidence for potential Ranablairi and the southern form of Leopard frog in gene flow between two species of spadefoot toads, Southeastern Arizona (Anura: Ranidae) SouthwestNatur. Scaphiopusbombifronsand Scaphiopushammondi. 22:443-453 ". Herpetologica 31:286-288. FROST, J. S. and J. E. PLA1Z. 1983. Comparative assess­ FROST, 1. S. 1982. Functional genetic similarity between ment of modes of reproductive isolation amoung four geographically seperated populations of Mexican leop­ species of leopard frogs (Rana pipiens complex). ard frogs (Ranapipiens complex). Syst. Zool. 31:57-67. Evolution 37:66-78.

FROST, J. S. and J. T. BAGNERA. 1977. Sympatry between

Southwestern Herpetologists Society SPADEFOOT TOAD HYBRIDIZATION 181 GARTSIDE, D. F. 1980. Analysis of a hybrid zone between LnTLEJOHN. 1978. Analysis of a narrow hybrid zone chorus frogs of the Pseudacrisnigrita complex in the between two species of Pseudophryne (Anura: southwestern United States. Copeia 1980:56-66. Leptodactyllidae) in South-Eastern Australia. Evolution 32:602-612. GERHARDT, H. c., S. 1. GlJITMAN and A. A. KARLIN. 1980. Natural hybrids between Hyla cinerea and Hyla MECHAM, 1. S. 1960. Introgressive hybridization between gratiosa: morphology, vocalization and electrophoretic two southeastern treefrogs. Evolution 14:445-457. analysis. Copeia 1980:577-584. MOORE, W. S. 1977 . An evaluation of narrow hybrid zones GOLLMAN, G. 1991. Population structure of Australian in vertebrates. Quart. Rev . Bio!. 52:263-277. frogs (Geocrina laeviscomplex) in a hybrid zone. Copeia 1991:593-602. NEWMAN, R. A. 1988a. Genetic variation for larval anuran (Scaphiopuscouchi) development time in an uncertain GREEN, D. M. 1981. Theoretical analysis of hybrid zones environment. Evolution 42:763-773. derived from an examination of two dissimilar zones of hybridization in toads (genus Bujo). Ph . D. Dissertation. _____ 1988b. Adaptive plasticity in development of Univ. of Guelph. Canada. Scaphiopuscouchi: tadpoles in desert ponds. Evolution 42:774-783. HARKEY, G.A. and R.D. SEMLITSCH. 1988 . Effects of temperature on growth, development and color PATION, J. L., J. C. HAFNER, M. S. HAFNER and M. F. polymophism in the ornate chorus frog. Copeia SMITH. 1979 . Hybrid zones in Thomomys bottae pocket 1988:1001-1007. gophers: genetic phenetic and ecologic concordance pat­ terns . Evolution 33:860-876. HARRISON, R. G. and J. ARNOLD. 1982. A narrow hybrid zone between closely related cricket species. Evolution PLATZ, J.E. 1981. Suture zone dynamics: Texas populations 36:535-552. of Rana berlandieri and Rana blairi. Copeia 1981:733­ 734 . HEWITI, G.M. 1988. Hybrid zones - natural laboratories for evolutionary studies. Trends Ecol. Evo!. 3:158-167. POMEROY, L. V. 1981. Developmental polymorphism in the tadpoles of the spadefoot toad Scaphiopus HilLIS, D. M. 1981. Postmating isolating mechanisms multiplicatus. Ph. D. Dissertation. University ofCalifor­ among three species of the Rana pipiens complex in nia, Riverside, California. Texas and Southern Oklahoma. Copeia 1981:312-319. RAND, D. M. and R. G. HARRlSON. 1989. Ecological HUGHES, N. 1%5. Comparison of frontoparietal bones of Genetics of a mosaic hybridzone: mitochondrial, nuclear Scaphiopus bombifronsand Scaphiopus hammondi as and reproductive differentiation of crickets by soil type . evidence of interspecific hybridization. Herpetologica Evolution 43 :432449. 21:196-201. REMMINGTON, C. L. 1968. Suture-zones ofhybrid interac­ LICHT, L. E. 1974. Survival ofembryos, tadpoles and adults tion between recently joined biotas. Evol . Biol. 2:321­ of the frogs Rana aurora aurora and Rana pretiosa 428. pretiosa sympatric in southwestern British Columbia. Can. 1. Zoo!. 52:613-627. RUIBAL, R., L. TEVIS Jr. and V. ROIG. 1969. The terrestrial ecology of the spadefoot toad Scaphiopus hammondii. MABRY, C. M. and J. L. CHRISTIANSEN. 1991. The Copeia 1969:571-584. activity and breeding cycle of Scaphiopus bombifronsin Iowa. J. Herpetology 25: 116-119. SATTLER, P. W. 1978. Biochemical genetic investigations of introgressive hybridization and systematic relation­ MARTOF, B. S. 1961. Vocalization as an isolating mecha­ ships in the spadefoot toads, genus Scaphiopus. Ph.D. nism in frogs. Amer. Midi. Natur. 65:118-126. dissertation, Texas Tech. University, Lubbock, Texas.

MAYHEW, W. W. 1965. Adaptations of the amphibian SATTLER, P. W. 1985 . Introgressivehybridization between Scaphiopus couchi, to desert conditions. Amer. Mid!. the spadefoot toads Scaphiopus bombifrons and Natur .74:95-109. Scaphiopus multiplicaius (Salientia; Pelobatidae),Copeia . 1985:324-332. McDONNELL, L. J., D. F. GARTIISIDE and M. J.

Herpetology of the North American Deserts 182 MARIE A. SIMOVICH SEMLITSCH R. D . and H. M. WILBUR. 1988 . Effects of WILBUR, H. M and 1. P. COLLINS. 1973. Ecological pond drying time on metamorphosis and survival in the aspectsofAmphibianmetamorpbosis. Science 182:130" salamander Ambystoma talpoideum. Copeia 1988-978­ 1314. 983. WILBUR, H. M., D.l. RUBENSTEIN and L. FAIRCHILD. SIMOVICH, M . A. 1985. Analysis ofa hybrid zone between 1978. Sexual selection in toads: the roles of female the spadefoot toads Scaphiopus multiplicaius and choice and male body size. Evolution 32:245 -263 . Scaphiopus bombifrons. PhD. dissertation. University of California, Riverside. WOODRUFF, D. S. 1973. Natural hybridization and hybrid zones. SYSl Zool. 22:213-218. SIMOVICH, M . A. and C. A. SASSAMAN. 1986. Four independent electrophoretic markers in spadefoot toads . WOODWARD, B. D. 1987. Interactions between 1. of Heredity 77:410-414. Woodhouse's toad tadpoles. Copeia. 1987:380-386.

SIMOVICH, M. A., C. A. SASSAMAN and A.CHOVNICK. 1991. Postmating selection of hybrid toads (Scaphiopus multiplicatus and Scaphiopus bombifrons). Proc. San Diego Soc. Natural History.

STEBBINS, R. C. 1985. Field Guide to Western Reptiles and Amphibians. Houghton Mifflin Co. Boston, Mass.

TEMPLETON, A. R. 1981 . Mechanisms of speciation - a population geneticapproach. Ann . Rev .Ecol. Sysl12:23­ 48.

THORNTON, W. A. 1955 . Interspecific hybridization in Bufo woodhousei and Bufo valliceps. Evolution 9:455­ 468.

TRAVIS, 1. 1980. Phenotypic variation and the outcome of interspecific competition in hylid tadpoles. Evolution 34:40-50.

TRAVIS, 1. 1983. Variation in development patterns oflarval anurans in temporary ponds.I. Persistant variation within a Hyla gratiosa population. Evolution 37:496-512.

TRAVIS, 1.1984. Anuran size at metamorphosis: Experi­ mental test ofamodel based on intraspecificcompetition. Ecology 65:1155-1160.

WASSERMAN, A. 0 .1957. Factors affecting interbreeding in sympatric speciesofspadefoot toads. Evolution 11:320­ 338.

WELLS, K. D. 1977 . The social behavior ofanuran amphib­ ians. Anim. Behav. 25:666-693.

WEINS, 1.1. andT.A. TITUS. 1991. A phylogenetic analysis of Spea (Anura: Pelobatidae). Herpetologica 47:21-28.

WILBUR , H. M. 1977 . Density-dependent aspects of growth and metamorphosis in Bufo americanus. Ecology: 196­ 200.

Southwestern Herpetologists Society