F. B. Browne 351 
-The food-material, in tlie form of resting larvae and of pupae of various 
species, was placed in the cell and every movement of the parasite on its host 
could be watched under a Zeiss binocular dissecting microscope. The tube 
could be rotated so as to cause the host larva or pupa to roll over so that 
Melittobia eggs and larvae could be seen and even counted, if not numerous, 
without removing the host from the cell. As a rule however the cell was opened 
and the host tipped out whenever a count of eggs and larvae of the parasite 
was to be made and at these times the parasite sometimes remained in the 
tube and sometimes came out upon the host. In the latter case she frequently 
remained upon the host the whole time that her eggs and larvae were beim; 
removed and usually showed no signs of fear or annoyance even when pushed 
out of the way by the brush or needle. 
Occasionally the cells required cleaning as the glass became dulled by the 
excreta of the female. 
The cotton-wool plugs seemed to permit the passage of all necessary air 
and in every way these cells were most convenient to handle. Most of my 
experiments were made with the larvae of species of Odynevus and of Osmia 
rufa but, as will be seen later on, the latter was, for some unknown reason, 
not an ideal host, so that most of my results were obtained from experiments 
with the wasp larvae. 
II. The Hosts of Melittobia. 
Melittobia is parasitic upon a number of kinds of bees and wasps. 
Howard (1892) gives a list of 14 different genera which are attacked by it. 
Malyshev (1911) adds three more and Girault (1912) adds another two. It 
was also known to be parasitic upon various flies of which it attacks the pupa. 
Experiment however shows that almost any larva or pupa of bee or wasp 
is attacked by it under laboratory conditions and even beetle larvae and pupae 
served as food for both imago and larva. 
But although almost any insect larva or spider in a quiescent condition— 
such as those paralyzed and stored up by various fossorial and Eumenid wasps, 
or larvae about to pupate—seemed to be regarded as suitable for oviposition 
and as food for the resulting larvae, in many cases such larvae failed to mature, 
mainly, I believe, because the food material dried up and the parasites con- 
quently starved. 
The following is a list of the genera of hosts previously recorded: 
Pison 
Sceliphron 
Chrysis 
M cmodontomerus 
Leucospis 
Musca and other fly pupae. 
Anthopliora 
Osmia 
Chalicodoma 
Iieriadcs 
A nthidium 
Bomb us 
Megachile 
Ceratina 
Prosopis 
Stelis 
Odynerus 
Trypoxylon 
Agenia 
23—2 
