8 BULLETIN 345, U. S. DEPARTMENT OF AGRICULTURE. 
Many attempts were made by Mr. Pomeroy to keep isolated pairs 
of fles under as natural conditions as possible in order to obtain not 
only the preoviposition period but also the number of depositions per 
fly and the number of eggs per deposition. The usual methodwas 
to remove from a lot of laboratory bred flies the first pair which were 
found in copulation. After transfer to a separate cage they were 
supplied with food and a medium for oviposition and were kept under 
observation in much the same way as the larger lots. The results of 
the tests with isolated pairs are summarized in Table IT. 
The fact that only 7 out of some 30 tests gave any positive results 
points to the difficulties in obtaining significant data from experi- 
ments performed under artificial conditions. 
The shortest preoviposition period was about two and one-half days 
(Table I, lot No. 1). Mr. Pomeroy’s original notes on this lot show 
that the ‘‘flies emerged some time between August 18, 10.30 a. m., 
and August 19, 9.30 a. m.,’’ at which time they were ‘‘ transferred to 
cage and supplied with banana, water, and fresh manure. Examined 
August 20 at 3.30 p.m. No eggs were found. Fresh manure sup- 
plied. August 21, at 1 p. m., two clusters of eggs were found, 80 to 
90 eggs in each cluster.”” Now, if the flies which deposited these eggs 
had emerged on the afternoon of the 18th the preoviposition period 
was a little longer than two and one-half days. If they had emerged 
during the morning of the 19th before 9.30 a. m. the period was a little 
less than two and one-half days. Other remarkably short periods 
are to be found in the three 3-day periods shown in the table, and in 
the four 4-day records. It is reasonable to suppose that if these short 
periods were found under artificial experimental conditions they 
would be not at all uncommon under natural conditions. 
‘! The longest period was one of 23 days (Table I, lot No. 69). In 
general it may be said that the shortest records occurred in mid- 
summer and that the longer ones were obtained during the autumn 
months. It is true that some long periods were found in summer, 
e. g., a 14-day and a 16-day record in July, but no short records were 
ever found during the autumn months. In other words, the length of 
the preoviposition period is greatly influenced by the temperature. 
‘A study of Table I will show that the arrangement according to the 
increasing length of the period corresponds in a rough way with the 
order of the decreasing mean temperatures of those periods. The 
relation between the temperature and the length of the preoviposition 
period is made clearer in the correlation table (Table IIT), in which 
the columns give the preoviposition period in days and the rows 
represent 2° intervals of temperature. 
