38 TRANSACTIONS LIVERPOOL BIOLOGICAL SOCIETY. 
the series as shown starting from the imago of the previous 
season to that of abundance. 
Ihave added the state the insect would be in during 
the various months of the season, and the figures at the 
bottom indicate the monthly means for a period of 50 
years. Now let me first confine attention to temperature, 
and I may as well at once admit that I can get no sort of 
satisfactory theory out of the temperature figures. 
The assumption would seem probable at first sight that 
temperature at the time of the flight of the imago might 
affect oviposition. Not at all. June and July are the 
imago months, and for the purpose of comparison I have 
singled out two other years in which there were no galu 
seen at all. Now the mean temperature of June, July is 
60°6 but in 1869 it was only 54°9 followed by a high average 
for June, July, 1870, while 1887 was high and 1888 very 
low: and each of these exactly opposite series results in 
a galw season, while 1868 almost the highest June, July 
of the century produced nothing. 
Similar results are obtained by a consideration of the 
effect of temperature on any other stage of the imsect 
career. August 1s the egg month, early September the 
young, and late September and October the mature 
larval period. The experience of most entomologists as 
regards larve is that they are extremely susceptible to 
changes of temperature, yet the larve of 1887 must have 
flourished during September and October months, both 
two degrees below the mean quite as much as in 1858 
when they were nearly four degrees above it. , 
From the first three series 1t might well be argued that 
a high summer temperature at any rate during the second 
half of the biennial term had a considerable effect ; but it 
stands to reason that similar influences work each year 
in a similar way, and it is part of my theory that the 
