2.92 T. H. Morgan 
The more important facts brought out by a study of the inhabi- 
tants of these galls remains to be discussed. Some few galls con- 
tain only male producers, and there is one record at least where all 
the migrants were female producers. The great majority of 
the galls contained both sorts of migrants, with a preponderance 
of male producers. ‘The results have a rather unique bearing on 
the question of sex-determination as will be pointed out later. 
Two additional facts of prime importance must be also noted 
here. I kept a record of the size of the galls. No relation between 
the proportion of male to female producers and the size of the 
gall was observed. In nearly all of the galls a few or even many 
very small individuals may be found. ‘These carry a very few 
eggs, but the eggs are full size of their kind. I have observed 
cases where only a single female egg wes present. ‘There can be 
little doubt that lack of food is the cause of the small size of these 
individuals. Their greater abundance amongst the later hatched 
individuals supports this interpretation as well as other facts 
observed. It follows, therefore, that food conditions do not 
determine whether a migrant becomes a male or a _ female 
producer. It is probable, therefore, that internal factors have 
already determined the result. Whether these internal factors 
can be traced to internal or to external conditions in the preceding 
generation is not evident from the facts given, but there is another 
point that would seem to indicate that the factors are internal in 
the first generation also. ‘The stem-mother continues to lay her 
eggs up to the time when the gall opens. Before this event when 
the gall is swarming with larve and the walls completely lined by 
them, the conditions for obtaining food must be less favorable 
than during the early stages of the gall. Yet both amongst the 
first eggs laid, and amongst the last laid, maleand female producers 
occur. It is difficult to see how food conditions determine the 
result; for even if the eggs should be affected only at a certain 
period on which the momentary condition of nourishment might 
act still we should expect a greater preponderance of one or of the 
other sex, according to whether the general conditions are favor- 
able or the reverse. 
The data furnish strong evidence in favor of some sort of 
