206 
Psyche 
[Vol. 90 
glandular secretions in overcoming the constraints of humidity. 
While Eickwort et al.’s (1981) hypothesis of nesting evolution in the 
Megachilidae is interesting, more recent evidence points to the fact 
that glandular secretions are important features of megachild nest- 
ing biologies. Parker and Tepedino (1982) observed the application 
of salivary secretions to bare walls of Osmia marginata Michener 
nests. Frohlich (1983) observed the incorporation by Osmia bruneri 
Cockerell of an abdominal secretion into the provision, and also 
noted the application of a salivary secretion to partitions. Dianthi- 
dium ulkei ulkei (Cresson) also incorporates an opaque viscous sub- 
stance, originating from the abdomen, into resinous cell walls and 
spreads the material over bare areas of the cell (Frohlich and 
Parker, in prep.). 
Since the twig nesting megachilids probably arose from soil nest- 
ing megachilids (Eickwort et al. 1981) and since the Megachilidae is 
distantly related to the other soil nesting families (Michener 1974) 
we propose that the Megachilidae have retained the habit of using 
glandular secretions to line cells. It seems likely that a waterproof 
layer of some sort is necessary to maintain the humidity of the cell 
within tolerable limits. Lining cells with leaves in soils that are 
particularly moist would have little effect on reducing humidity and 
controlling fungal growth nor would lining pre-existing cavities 
such as twigs prevent dehydration. It is important, therefore, that 
we thoroughly examine the behavioral, and more importantly chem- 
ical, components of nesting in order to gain an understanding of the 
role of nest architecture in evolution. 
The paucity of information available on the nesting biologies of 
other species of Eumegachile make it difficult to confirm (or, refute) 
Mitchell’s (1981) recent revision of the old genus Megachile on the 
basis of behavior or nest architecture. The available data does con- 
firm the separation of Eumegachile from Megachile, since the nest 
architectures of the two are radically different. Eumegachile pug- 
nata nests are similar to, though somewhat more elaborate than, 
known nests of other species in the subgenus. Krombein (1967) 
reported that E. inimica inimica Cresson makes unlined cells with 
partitions of agglutinated sand a little larger than the inner circum- 
ference of the nest. Eumegachile inimica sayi Cresson also uses a 
single leafcutting as a partition but covers it with leaf pulp, incorpo- 
rating five pebbles (Krombein 1967). Eumegachile policaris Say lays 
