HOST SELECTION BY EGG PARASITES
The nature of the oviposition stimulus(i) for the wasp parasites of cockroach eggs is unknown. Edmunds (1954) noted that Prosevania punctata showed more interest in oöthecae that had been cemented to the substrate than in clean oöthecae that had simply been dropped. Cros (1942) experimented with two females of P. punctata to see if the wasps could find oöthecae that had been buried in sand by the oriental cockroach. After prospecting the sand with their antennae, the wasps dug deep excavations with their front legs but always mistook the location of the oöthecae. Cros suggested that the wasps were misled by the odor left in the jar by the cockroaches. It is quite possible that odor helps the wasp find the host oötheca.
The extent of host selection varies among these parasites; some species will oviposit into the eggs of more than one species of cockroach, but others show some degree of host specificity. Positive selection of specific hosts by certain parasites appears in correlative data from different investigators on pages 235 to 254. There is a small body of data that shows nonacceptance of certain hosts by some of these wasps. For example, Comperia merceti would not parasitize eggs of Blatta orientalis or Periplaneta americana in the laboratory (Lawson, 1954a). We (unpublished data, 1957) exposed a soft oötheca, recently removed from Eurycotis floridana, to C. merceti; no wasps developed; we had similar negative results with C. merceti and oöthecae of B. germanica. We (1954b) could not induce Tetrastichus hagenowii to parasitize eggs of Blattella germanica, B. vaga, or Parcoblatta virginica in the laboratory. In our experiments, T. hagenowii oviposited into eggs of Supella supellectilium, but the wasp eggs either failed to hatch, or if they hatched, the larvae died before completing development. Neither would T. hagenowii parasitize eggs of N. rhombifolia (Roth and Willis, unpublished data, 1957). Anastatus tenuipes would not parasitize the eggs of Latiblattella lucifrons Hebard, Periplaneta americana, B. germanica, or B. vaga (Flock, 1941). Anastatus floridanus would not oviposit into eggs of S. supellectilium and only rarely into eggs of P. americana or B. orientalis (Roth and Willis, 1954a); in the laboratory, this wasp could not be maintained beyond one generation on the eggs of P. americana. Edmunds (1953b) could not induce Prosevania punctata to parasitize eggs of B. germanica. Cros (1942) induced P. punctata to oviposit into a mantid oötheca, but neither mantids nor parasite developed.