BUCKINGHAM. — DIVISION OF LABOR AMONG ANTS. 449 



To eliminate as far as possible other factors than that of food, the 

 latter was placed alternately in chamber B and in chamber A, the same 

 number of observations being made for each position. 



2. Observations. — In these experiments ten different colonies were 

 made use of llccords in regard to the feeding activity were made 

 every 24 hours with the exception of certain Sundays and a few holi- 

 days. It is clearly necessary, however, to exclude observations made 

 on any day succeeding one when no observations had been made, 

 because the time interval during which the ants had had an oppor- 

 tunity to respond was in such a case 48 hours instead of 24 hours, the 

 usual interval. At the beginning of each observation the ants were 

 all placed in chamber A, and the food was placed half the time in 

 chamber .1, half the time in chamber B. A ''series" of observations 

 consists of a number (50 or 25) of separate observations all made under 

 the same conditions as to the relation of insects to the food-chamber. 

 The maximum number of observations in a week was only five, except 

 during those periods when records were made on Sundays as well as 

 week days, in which cases seven observations a week were possible. 

 During the earlier experiments on this species, 50 observations were 

 included in each series, but later the number was reduced to 25. A 

 "set" of observations includes two or more "series," one of the series 

 being made while the food was in chamber A ; the other while it was 

 in B. The ants at the beginning of every experiment were, as stated, 

 all in one chamber {A). After the lapse of 24 hours the number of in- 

 dividuals which had migrated into chamber B was noted. The results 

 of a series of 50 (or 25) such observations were combined as follows : 

 First was computed what per cent of the whole number of ants under 

 observation made their way into the unoccupied chamber during each 

 observation. These per cents for the 50 (or 25) observations were then 

 averaged. The result is shown in Table I. When the food was in 

 chamber B, these values give a partial measure of the stimulus to mi- 

 gration caused by food ; but, assumably, the migration under this con- 

 dition was not due exclusively to the stimulus of food. To eliminate 

 as far as possible all other factors except food, experiments were also 

 made with the food in chamber A. The individuals making their way 

 from chamber A to chamber B under this condition would clearly not 

 be attracted to B by food, and the number of them may fairly be taken 

 as an approximate measure of the number of individuals which made 

 their way into the food-chamber (when that chamber was B) indepen- 

 dently of the influence of food. In the earlier sets of observations (viz. 

 with 50 observations in a series) three series were combined into a set 

 (colonies 4, 17, 20, and 24) ; in the later sets only two series (colonies 



VOL. XL VI. — 29 



