Townsendiella rufiventris Linsley, 1942

Orr, Michael C. & Griswold, Terry L., 2015, A review of the cleptoparasitic bee genus Townsendiella (Apidae, Nomadinae, Townsendiellini), with the description of a new species from Pinnacles National Park, ZooKeys 546, pp. 87-104 : 96-97

Taxon classification Animalia Hymenoptera Apidae

Townsendiella rufiventris Linsley, 1942 Fig. 3b

Holotype.

female, pinned; Palm Springs, California; Mar 26, 1932; deposited in CAS (Type#14881).

Diagnosis.

The female of this species may immediately be separated from the other Townsendiella by the presence of the lunule on T5 (apicomedial impressed rim with dense, fine punctures). Both females and males also have a strong medial production on the metanotum, which is not seen in other species. The male gonoforceps are quite distinctly flattened and relatively transparent, lacking a distinct gonostylus.

Distribution.

The distribution of Townsendiella rufiventris is exceptionally broad among the Townsendiella , spanning from Baja California, Mexico, and the eastern Sonoran Desert, extending northward through the coastal ranges nearly to the San Francisco Bay. As such, this species has both the most southerly and westerly collection events of any Townsendiella . It inhabits several ecoregions, given here in order of most to least collection localities: Mojave Desert, Sonoran Desert, California coastal sage and chaparral, California interior chaparral and woodlands, California montane chaparral and woodlands, and Baja California Desert.

Phenology.

This species is known to fly from mid-March through July, although its phenology appears to differ throughout its range. Within the Mojave, it appears to be most active from late March through May. In the South Coast Range, however, it appears to be active in June and July. Further collections from its northern distribution are necessary to test this possibility.

Bee hosts.

Interestingly, Townsendiella rufiventris appears to use the halictid genus Conanthalictus Cockerell as hosts, although prior publications have not listed hosts at the species-level ( Linsley 1958, Rozen and McGinley 1991). A determination label by the late Paul D. Hurd with a date of 1963 from the Essig Museum gives the determination of Townsendiella rufiventris and states it was "flying about nest site of Conanthalictus nigricans Timb." The label was placed before a series of Townsendiella rufiventris from "San Marcos Ranch HQ, Santa Inez Mts," found near Santa Barbara Co., California. More recently, Townsendiella rufiventris has been collected northwest of San Bernardino, California ("N Sierra Ave") invading the nests of Conanthalictus bakeri Crawford, 1907 (D. Yanega, unpublished observations, 30 April 2015). Circumstantial evidence has also been found in the association of high numbers of Townsendiella rufiventris at sites with Conanthalictus bakeri ("Jamul CA" and "Spring Valley CA") and Conanthalictus wilmattae Cockerell, 1936 ("Anza-Borrego, In-Ko-Pah Park" and "Anza-Borrego: Peña Spring"), all sites which yielded few to no Hesperapis , though no positive host associations were possible (J. Hung, unpublished observations, 14 May 2015).

Floral hosts.

Asteraceae : Lasthenia californica DC. ex Lindl.; Boraginaceae : Cryptantha intermedia (A. Gray) Greene, Cryptantha sp. Lehm. ex G. Don, Phacelia distans Benth., Phacelia sp. Juss.; Onagraceae : Chylismia munzii (P.H. Raven) W.L. Wagner & Hoch.

Discussion.

The possibility that Townsendiella rufiventris is two species was explored based on observations by Doug Yanega (pers. comm., 10 December 2013). The primary character investigated was the form of the lunule on the female T5, a finely-pitted, tessellate apicomedial depression filling the otherwise concave rim, the presence of which is a unique character for Townsendiella rufiventris . The species may be roughly split into two series, those with a flat apical rim on the lunule (series 1) and those with an apical lunule which projects farthest medially (series 2; allied with type of Townsendiella rufiventris ). However, the reliability of the lunule as a character is questionable in light of its apparent flexibility, demonstrated by the variability in its angle relative to the rest of T5 across specimens of the same series. The area basal to the lunule is also variable, going from sparsely pitted and shiny in series 1 to densely, craggily pitted and dull in series 2, although numerous exceptions have been discovered. The proportion of black integument basal to the lunule also varies, with more in series 1 and less in series 2, but exceptions to this have also been found. The male genitalia were also examined, using four males collected with females of series 1 and three specimens associated with series 2 females, but no diagnostic characters were detected. No characters from either sex which clearly and consistently delineate the two entities were discovered. A geographic split is also impossible; although series 2 is primarily found in southern California, the range of series 1 appears to completely envelop that of series 2. It must also be noted that females of both series 1 and series 2 were found from the same collection event thrice, casting further doubt on the existence of two species.