400 МВ. С. Е. М. SWYNNERTON ОМ 
in No. 11, but, at any rate on one side, close-packed as in No. З. 16 ‘acci- 
dentals.’ 
* No. 19. A very good raceme, but barely beginning to make a mat. 
Barely 100 flowers, dense and not very conspicuous individually—much like 
№. 3. 24 ‘accidentals.’ 
* No. 20. A large lax raceme, with flowers rather large-mouthed as in 
No. 11 but by no means tightly packed. Still, only one definite fault, and 
that barely coming within the region of the present flowers. About 150 
flowers out, and only 5 ‘ accidentals.’ 
* No. 21. Of exactly the same type as the last. 69 flowers out, no fault, 
only one ‘accidental.’ 
* No. 22. About 130 flowers, same type as last, no faults; 2 ‘accidental ? 
slits and 2 or 3 bill-point marks. 
“The total number of flowers given is in each case only approximate, as I 
found it difficult to draw an exact line between flowers still in use and those 
that were already, by being bruised and torn by the bird's feet, becoming 
consigned to the * mat.’ 
“TI again watched hive-bees at the pollen and honey and a smaller bee at 
pollen: also pollen-eating flies. Ants of at least three species were also 
present. 
“Visited yesterday's two Kniphofia on the outskirts. That damaged 
yesterday had been weighed down half-way to the ground by something but 
showed no other damage. The other (undamaged yesterday) was to-day 
damaged as the first had been yesterday. Fewer flowers were implicated 
(about 7 in one place and 1 in another) and the work was neater. Оп 
yesterday's I found a very good piece of evidence (previously overlooked) 
to indicate that it had been the work of a bird. Three flowers had been. 
seized together, two of them showed the clear impress of the upper mandible 
and one of the lower, as shown in РІ. 82. fig. 3. 
“Оп leaving I removed some of the flowers from the upper surface of 
the leaning head in order to see whether the flowers thus exposed would be 
pierced.” 
“June 6th. Examined the day before yesterday, above the dam, seven 
or eight heads of Kniphofia in which the flowers had reached the apex and 
could therefore be attacked from above, even if awkwardly, by sunbirds 
perching there. About half showed no damage at all other than the usual 
slits ; the remaining heads did show, in a small proportion of only their 
uppermost flowers, definite punctures and slits on their upper surfaces. 
These might have been made by a bird perching either above or below, and 
in either case would probably have enabled him to avoid contact with the 
anthers. In one ease, а raceme with about 50 flowers open, the apex was 
bent over in such a way as to afford quite a convenient perch to a bird, 
and a large proportion of the flowers growing near the apex were slit 
