|7 « 
to ehano** It will probably be neceset ry to col loot further 
observations to refute this possibility. 
Since the broodlnees of bird© also Increases towards 
the time of hatching I wits In to res tel in the relation between till© 
and swooping. -In the first ?&000 I bad to try and §•• If I 
could dMoaotrote snob an lacrs&ee in broodlaees* raver ■ 1 waya*^ 
attempted but the following on© showed west elc&rly th© effect. 
During the tine of tnoub© tion occasional panics* five or six 
during a Doming, were still disturbing th« colony and each 
on® offered a choice# ao id i i ■ v , to each br ling bird 
whether It should leave Its nest or not, Stowe tin## a bird 
stayed and sometimes it left. One would expect that the oh nose 
of staying increased the more broody the bird was, I made noted 
of the choices at each panic for several neat© which were 
often si different stages of incubation* Then at the end I 
collected the datafrai all the nests for each 'day o€ the InoubatloA 
period* 'For instance rutting together all the mosti t isr* we re 
*>i choices made by birds on the nth day of Incub&tion, In only 
if of the® ( IJd) had the brooder stayed on the nest. To ®a’*© 
the percentage oalcula t -d acre valid I grouped the days in 
pairs {or longer periods when the number of observations 
Was few)* The following histogram shows that the percentage does 
in fact Increase slightly towards hatching and Is specially 
high Just after the young, hatch* A test of the data In the first 
and second halves of incube i m suggest that there le 
a real increase even before the egg© begin to chip* 
