32 BULLETIN 1222, U. S. DEPARTMENT OF AGRICULTURE. 
ing is being resumed. Further evidence to substantiate this conclu- 
sion is found both by weighings of the food present and by a com- 
parison of the composition of food found in cells containing 2-day 
and 3-day larvae. Sturtevant shows that there is a considerable 
increase in the amount of food found in cells containing 3-day larva? 
over the amount found with 2-day larvae. The composition of the 
food is changed also by the third day. This is most conspicuous 
through the fact that the color of the food found within the cells 
containing larvae of 3 days or more varies with the color of the 
pollen being brought in from the field, no stores of old pollen being 
present in the hive used in this experiment. The larval food found 
in cells containing 1-day and 2-day larvae is of a uniformly grayish- 
white color. 
With the change of food it appears that a different method of 
feeding is gradually adopted, mass feeding giving place to pro- 
gressive feeding; that is, the food is no longer given in quantity 
and in excess of immediate needs, but is supplied at about the same 
rate at which it is ingested. The bee larva is a rapidly growing 
organism, hence it follows that if the food is supplied from time to 
time, it must be introduced in rapidly increasing amounts. This 
necessitates either more visits or longer feeding periods, or both of 
these factors may be increased. The last is actually the case. The 
number of visits is increased, particularly on the fourth and fifth 
days, and the time of the visits is notably lengthened. 
Mass feeding, if this is strictly what happens in the case of the 
food first supplied, taken together with the later progressive feed- 
ing, shows an interesting similarity to the behavior of certain soli- 
tary bees y some of which practice mass feeding exclusively, while 
others, recently investigated, practice progressive feeding. It also 
lias a resemblance to the feeding behavior of the bumblebee, which 
lays the eggs on a lump of " pollen-paste " to furnish the first food 
after hatching, and then some time after the larva is hatched a 
type of progressive feeding is begun. A similarly modified type of 
mass feeding is shown in the case of the queen larvae of the honeybee. 
The large queencell is rapidly crowded with food before it is 
sealed, and the larva completes its development while eating more 
food which is inclosed with it. Frequently a mass of dried food 
remains even after the emergence of the queen, which confirms the 
supposition of mass feeding in this instance. 
The fact that the food of the larva up to 2 days is placed in the 
cell beside the larva and is apparently not fed into the mouth seems 
to preclude the possibility of mutual feeding of the nurse bee and 
larva during this period. Such reciprocal feeding has been de- 
scribed for certain social wasps and ants, and to this phenomenon 
Wheeler has given the name trophallaxis. Since the queen larva is 
also fed by mass feeding, perhaps exclusively by that method, there 
is little or no indication of trophallaxis in this instance. Roubaud 
has shown that in certain wasps in which trophallaxis is observed 
there is a decided disproportion between the amount of material 
given by the nurse wasp and that received by it from the larva being 
fed, the secretion received being sometimes greater than the food 
given by the nurse. Reciprocal feeding has never been observed in 
the honeybee, but since the food given to the older larvae is not com- 
plex in character and probably requires no great amount of work on 
