Members of the Crotalus durissus species complex are widely distributed from Mexico to Argentina in areas with mainly seasonally dry tropical deciduous forest. Although four species (C. culminatus, C. durissus, C. simus and C. tzabcan) are currently recognized, species limits remain to be tested. Previous genetic studies suggest that C. durissus and C. simus may be paraphyletic and that at least one cryptic species may be present. We analyzed 2596 bp of DNA sequence data from three mitochondrial and one nuclear gene to infer phylogenetic relationships in the Neotropical rattlesnakes. We also examined museum and wild specimens as well as captive animals to analyze morphological characters. Our results suggest that the current taxonomy of the Crotalus durissus species complex does not reflect evolutionary history. We found strong support for five independent lineages within Crotalus simus (sensu lato), with genetic and morphological evidence for three previously recognized taxa and two new species, as well as three major lineages within C. durissus that each represent species hypothesis to be tested with additional evidence. We also found support to retain C. totonacus in the Crotalus molossus species complex. We suggest conservative taxonomic changes to the complex and related species, but more evidence is needed (e.g., morphology, ecology and venom composition) to clarify relationships among species.

Ontogenetic changes in venom composition have important ecological implications due the relevance of venom in prey acquisition and defense. Additionally, intraspecific venom variation has direct medical consequences for the treatment of snakebite. However, ontogenetic changes are not well documented in most species. The Mexican Black-tailed Rattlesnake (Crotalus molossus nigrescens) is large-bodied and broadly distributed in Mexico. To document venom variation and test for ontogenetic changes in venom composition, we obtained venom samples from twenty-seven C. m. nigrescens with different total body lengths (TBL) from eight states in Mexico. The primary components in the venom were detected by reverse-phase HPLC, western blot, and mass spectrometry. In addition, we evaluated the biochemical (proteolytic, coagulant and fibrinogenolytic activities) and biological (LD₅₀ and hemorrhagic activity) activities of the venoms. Finally, we tested for recognition and neutralization of Mexican antivenoms against venoms of juvenile and adult snakes. We detected clear ontogenetic venom variation in C. m. nigrescens. Venoms from younger snakes contained more crotamine-like myotoxins and snake venom serine proteinases than venoms from older snakes; however, an increase of snake venom metalloproteinases was detected in venoms of larger snakes. Venoms from juvenile snakes were, in general, more toxic and procoagulant than venoms from adults; however, adult venoms were more proteolytic. Most of the venoms analyzed were hemorrhagic. Importantly, Mexican antivenoms had difficulties recognizing low molecular mass proteins (<12 kDa) of venoms from both juvenile and adult snakes. The antivenoms did not neutralize the crotamine effect caused by the venom of juveniles. Thus, we suggest that Mexican antivenoms would have difficulty neutralizing some human envenomations and, therefore, it may be necessary improve the immunization mixture in Mexican antivenoms to account for low molecular mass proteins, like myotoxins.

Rattlesnake venoms may be classified according to the presence/absence and relative abundance of the neurotoxic phospholipases A ${}_2$ s (PLA ${}_2$ s), such as Mojave toxin, and snake venom metalloproteinases (SVMPs). In Mexico, studies to determine venom variation in Mojave Rattlesnakes (Crotalus scutulatus scutulatus) are limited and little is known about the biological and proteolytic activities in this species. Tissue (34) and venom (29) samples were obtained from C. s. scutulatus from different locations within their distribution in Mexico. Mojave toxin detection was carried out at the genomic (by PCR) and protein (by ELISA) levels for all tissue and venom samples. Biological activity was tested on representative venoms by measuring LD ${}_{50}$ and hemorrhagic activity. To determine the approximate amount of SVMPs, 15 venoms were separated by RP-HPLC and variation in protein profile and proteolytic activity was evaluated by SDS-PAGE (n = 28) and Hide Powder Azure proteolytic analysis (n = 27). Three types of venom were identified in Mexico which is comparable to the intraspecific venom diversity observed in the Sonoran Desert of Arizona, USA: Venom Type A (∼Type II), with Mojave toxin, highly toxic, lacking hemorrhagic activity, and with scarce proteolytic activity; Type B (∼Type I), without Mojave toxin, less toxic than Type A, highly hemorrhagic and proteolytic; and Type A + B, containing Mojave toxin, as toxic as venom Type A, variable in hemorrhagic activity and with intermediate proteolytic activity. We also detected a positive correlation between SVMP abundance and hemorrhagic and proteolytic activities. Although more sampling is necessary, our results suggest that venoms containing Mojave toxin and venom lacking this toxin are distributed in the northwest and southeast portions of the distribution in Mexico, respectively, while an intergradation in the middle of both zones is present.