Melpomene sp.

Pape, Robert B., 2024, Biology and ecology of a deep cave nesting spider wasp, Ageniella evansi Townes, (Hymenoptera: Pompilidae), in Arizona, Journal of Natural History 58 (29 - 32), pp. 963-1054 : 999-1003

Melpomene sp. O. Pickard-Cambridge, 1898

Agelenidae is a large family with currently in excess of 1300 valid species (WSC 2022). These mostly nocturnal spiders build non-sticky, funnel webs in a variety of habitats, including caves ( Bennett and Ubick 2005). Melpomene is a New World group, distributed from southern Arizona and northern Sonora to western Texas and south to Panama ( Roth and Brame 1972). Melpomene sp. ( Figure 20D ) is apparently rarely taken by A. evansi in the Park, with only two records, representing one half of one percent of total observed wasptransported host spiders. It seems reasonable that use of this spider represents opportunistic capture by the wasps.

No field surveys were conducted for any of the spider host species, so no information is available about their phenology or population dynamics. The primary host (by percentage, Selenops ) is difficult to census due to their cryptic and inaccessible habits in cracks, fissures and caves. The information used in this analysis is based solely on the observed spiders returned to the cave by the wasps. The distribution by month of the total observed spiders taken by the Arkenstone Cave population of A. evansi is shown in Figure 23 . The total (N = 404) includes all identified and 18 unidentified spiders observed between 1992 and 2020. The earliest spider capture record for A. evansi in the Park is 7 November (1998; Gnaphosidae ) and the latest capture record is 11 May (2003; Selenops sp. ).

An analysis of host type by month ( Figure 24 ) reveals exclusive early season (November and December) use of gnaphosid spiders. No Selenops were observed taken in November or December, although a total of only 12 spiders were recorded for those 2 months. Use of gnaphosid spiders remains dominant through February. Use of Selenops overlaps gnaphosid use beginning in January, becomes dominant in March and extends through early May, while gnaphosid use drops off significantly by the end of February. Because Selenops apparently live more than 1 year ( Crews et al. 2008), various age cohorts are probably present throughout the year as potential hosts for A. evansi . This study revealed that A. evansi uses a variety of age cohorts of Selenops , and the lack of their use of this species early in the wasp active season suggests a preference for gnaphosid spiders. The drop-off in use of gnaphosid spiders may be the result of a depletion of large individuals of this element of the wasp host base within the wasp foraging range by the end of February. Switching of host type in the Pompilidae resulting from depletion of certain elements of the host base has been documented elsewhere ( Kurczewski and Kurczewski 1968a; Martins 1991; Endo and Endo 1994). The wasps then apparently focus on Selenops for the remainder of the active season. The wasps take lycosids beginning in February, peaking in April and continuing into early May. Approximately one-quarter of spiders taken in April are lycosids. The use of lycosids beginning in February may reflect the time at which these spiders reach an acceptable minimum size threshold for the wasps.

An analysis of host type by year ( Figure 25 ) shows a greater diversity of spiders used in the early years of the study (1996–1997), where both lycosid and agelenid spiders show up in the data. This may be due to the greater number of spiders returned to the cave in years when the female hunting population was large. Lycosid and agelenid spiders are probably regular components of A. evansi hosts taken in a typical year, but are less often observed as they are apparently minor elements of total spiders taken. The broader implication is that when larger data sets are obtained for other pompilid species a greater diversity of host use may be found. Thus the Pompilidae , in general, will likely be found to be more flexible in their host selection that is currently recognised.

Spider size is apparently the crucial element in host selection, even among pompilid species that take spiders of more than one kind ( Endo and Endo 1994; Kurczewski and Edwards 2012). Kurczewski and Kurczewski (1968a, 1968b) provide abundant data supporting the use of large and/or adult spiders by larger pompilid species and smaller and/or immature spiders by smaller pompilid species. Ageniella evansi generally takes spiders that are subequal to or slightly shorter than their body length. Spiders much longer or larger than the wasps quickly become cumbersome and inefficient to haul long distances to the nesting site. This is particularly important for these cave nesting populations since once they enter the cave they still have a considerable distance to go before reaching their nest site. Whether or how this may affect the extent of their epigean hunting territory is not known.

The length of incoming spiders was estimated using a combination of the visual length of the wasp and the proportional length of the spider. Only the small handful of spiders that were separated from female wasps were physically measured, and are included in the data. The combined average length of all spider hosts observed was 9.2 mm (N = 311). The average length of spiders by type was: Gnaphosidae 9.7 mm (N = 76), Selenopidae 9.3 mm (N = 208), Lycosidae 7.6 mm (N = 25) and Agelenidae 8.5 mm (N = 2). The slightly higher average length in gnaphosid over Selenops spiders does not seem of a magnitude that alone would explain the apparent preference for gnaphosids. Proportionally, however, gnaphosids have abdomens of greater mass and this could conceivably be a driving factor in preference for those spiders. Since the wasps probably select for larger spiders, the noticeably smaller average length of lycosids may suggest these are smaller species than the other hosts used. Body length measurements for female wasps averaged 11.2 mm (8– 14 mm; N = 1548) and males 9.0 mm (6–10 mm; N = 347). Thus, on average, the female wasps are approximately 22% longer than the spiders they take. The group of spider host types selected by A. evansi suggests hunting of spiders is based on guild and/or hunt niche specificity rather than host species fealty (Evans 1953; Rodriguez et al. 2016).

Plotting annual percentage of hosts by type against the size of the yearly female hunting population shows the proportion of host types used remains consistent when the annual female population is approximately 19 or greater ( Figure 26 ). Smaller annual wasp populations do not provide enough spider return data to show whether host type distribution is consistent at those levels.

Figures

Figure 20. The four families of spider hosts taken by Ageniella evansi Townes, 1957 in Colossal Cave Mountain Park, Arizona. A, Selenopidae Selenops sp. Latreille, 1819. B, Gnaphosidae (Herpyllus schwarzi Banks, 1901). C, Lycosidae. D, Agelenidae (Melpomene sp. O. Pickard-Cambridge, 1898).

Figure 23. Monthly total observed spider hosts (N = 404) taken by the Arkenstone Cave population of Ageniella evansi Townes, 1957 between 1992 and 2020. The total includes all identified and 18 unidentified spiders. The average peak of the wasp hunting season over 32 years (1993 to 2024) is 5 March.

Figure 24. Percent host type by month taken by the Arkenstone Cave population of Ageniella evansi Townes, 1957 between 1992 and 2020.

Figure 25. Percent of observed host type by year taken by the Arkenstone Cave population of Ageniella evansi Townes, 1957 between 1993 and 2020.

Figure 26. Total annual host utilisation, percent by type, relative to maximum annual female population of Ageniella evansi Townes, 1957 at Arkenstone Cave (composite data from 1993–2020).