H. Bradley Shaffer

Evolution in a paedomorphic lineage. I: An electrophoretic analysis of the Mexican ambystomatid salamanders

Paedomorphosis, or the retention of ancestral juvenile features in descendant adults (Gould, 1977), is one of a general class of heterochronic events which has recently been invoked as a possible mode of evolutionary change. Gould (1977) and Alberch et al. (1979) among others have reiterated the arguments of Goldschmidt (1940) that it is simpler to change morphology through shifts in the timing of established developmental events than by incorporating mutations that may disrupt the developmental status quo. Although a much more thorough understanding of the genetics of such developmental shifts is necessary before arguments on their evolutionary flexibility can be evaluated, paedomorphosis does seem to be an important mode of morphological evolution (see Gould, 1977, for many zoological examples, and Guerrant, 1982, for a botanical case study). The importance of paedomorphosis in salamander evolution has long been recognized (Noble, 1931; Tihen, 1958; Wake, 1966), and represents one of the major ways in which new morphological characters are acquired in these amphibians. Paedomorphosis has been invoked to explain the evolution of single elements (e.g., the nasal bone in Rhyacotriton olympicus, Wake, 1980a), as well as the simultaneous change of many aspects of salamander morphology (e.g., Wake, 1966; Alberch and Alberch, 1981). The most striking example of the complete reorganization of the adult morphology attributable to paedomorphosis occurs when salamanders fail to metamorphose to the terrestrial condition and mature while retaining virtually the entire larval morphology throughout life. I term this phenomenon larval reproduction. (Pierce and Smith, 1979; Sullivan, 1980; and Wake, 1980b discuss this terminological point.) Operationally, reproductive larvae retain external gills and an enlarged tail fin, lack movable eyelids, and are obligatorily aquatic as sexually mature adults. Although larval reproduction is found, at least occasionally, in most families of salamanders, it reaches its greatest variability of expression in the Ambystoma tigrinum-mexicanum-dumerilii-Rhyacosiredon complex of the family Ambystomatidae (Tihen, 1958). This complex includes about 17 currently recognized species, distributed across North America from southern Canada to the southern rim of the Mexican plateau (Gehlbach, 1967). The group reaches its greatest taxonomic and metamorphic diversity at its southern extreme, in the Mesa Central of central Mexico. Fifteen of the 17 species are endemic to this region (Smith and Taylor, 1948), and individuals exhibiting larval reproduction are known from well over half of these taxa. Of these, the axolotl, A. mexicanum, has been widely cited as a prime example of paedomorphosis in vertebrate evolution (deBeer, 1958; Gould, 1977). Before the importance of metamorphic failure in ambystomatid evolution can be addressed, a phylogeny of the group must be established as a framework for examining the evolution of metamorphosis. Here, I use starch-gel electrophoresis to evaluate the phylogenetic history of the Mexican ambystomatid salaman-

PATTERNS OF VARIATION IN AQUATIC AMBYSTOMATID SALAMANDERS: KINEMATICS OF THE FEEDING MECHANISM

Patterns of variation in the feeding mechanism of three species of ambystomatid salamanders (Ambystoma dumerilii, A. mexicanum, and A. ordinarium) were studied to provide insight into the nature of variation in kinematic parameters of the jaw mechanism associated with prey capture. A nested analysis of variance design provided an assessment of the amount of variation in six kinematic variables (measured from 200 frames/sec films of feeding behavior) both among species and among individuals within species. For all six variables, a highly significant proportion of the variance was explained at the intraspecific level. Among species, the most robust discriminators were variables associated with movement of the hyoid. The variables reflecting gape and lifting of the head provided no significant discrimination among species and had large error variances. The hyoid apparatus is the most phylogenetically conservative component of the feeding mechanism in lower vertebrates and was the most stereotyped component of feeding behavior within the salamander species studied here.

The Relationship Between Allozyme Variation and Life History: Non-transforming Salamanders are Less Variable

Several recent papers have reported low levels of genetic variation in nontransforming, paedomorphic urodeles. Here, we report the results of an electrophoretic survey of 21 allozyme loci in two populations of the Ozark Hellbender, Cryptobranchus alleganiensis, confirming the extraordinarily low levels of genetic variation previously reported in a different laboratory. We then summarize the literature on 102 species of salamanders to compare genetic variation in transforming and non-transforming taxa. On average, non-transforming salamanders are significantly less variable than transforming ones, supporting the idea that low levels of allozyme variation in Cryptobranchus are associated with larval reproduction. We interpret this association between life history mode and genetic variation as reflecting a fundamental difference in levels of population extinction and recolonization between aquatic and terrestrial species. N 1977, Merkle et al. reported the results of an extensive survey of electrophoretic variation in the non-transforming salamander Cryptobranchus alleganiensis. Their survey demonstrated an extraordinarily low level of genetic variation: among the products of 24 loci examined from a total of 137 specimens representing 12 geographic samples from the four major river systems inhabited by hellbenders, only two loci showed any genetic variation, and nine of the 12 populations were completely monoallelic. Merkle et al. (1977) also pointed out that several other non-transforming salamanders (the mudpuppy, Necturus maculosus; the dwarf siren, Pseudobranchus striatus; and the twotoed amphiuma, Amphiuma means) all appear to have low levels of genetic variation and large amounts of nuclear DNA, and suggested a possible causal relationship. A prediction from the work of Merkle et al. (1977) is that there should be a reasonably strong correlation between life history variation (that is, metamorphosis vs larval reproduction) and genetic variation. However, the relationship between life history pattern and genetic variation has never been evaluated systematically. In this paper, we first report the results of a reexamination of genetic variation in two additional populations of the hellbender, C. alleganiensis, confirming the results of Merkle et al. (1977) in a different laboratory using a somewhat different set of loci. To interpret these results, we present a literature survey of electrophoretic variation in over 100 species of salamanders from six families and 19 genera. We conclude that there is a strong relationship between larval reproduction and genetic variation: larval reproducers are less variable, on average, than metamorphosing salamander species. However, we interpret this relationship as a difference in the population structures of transforming and non-transforming salamanders, rather than representing a causal relationship between structural genetic variation and components of fitness as has been previously suggested (Pierce and Mitton, 1980; Mitton and Grant, 1984).

Biosystematics of Ambystoma rosaceum and A. tigrinum in Northwestern Mexico

cyprinodontid fishes of the genus Fundulus, with the description of Fundulus persimilis from Yucatan. Occ. Pap. Mus. Zool. Univ. Mich. 568. , AND C. H. LOWE. 1964. An annotated checklist of the fishes of Arizona, p. 133-151. In: The vertebrates of Arizona, C. H. Lowe (ed.). Univ. Arizona Press, Tucson, Arizona. MINCKLEY, W. L. 1973. Fishes of Arizona. Arizona Game and Fish Dept., Phoenix, Arizona. PARENTI, L. R. 1981. A phylogenetic and biogeographic analysis of cyprinodontiform fishes (Teleostei, Atherinomorpha). Bull. Amer. Mus. Nat. Hist. 168:335-557.