436 Mac far lane— Observations on Pitcher ed 
to be so invariable in certain species as to suggest that the 
cavity resulted from epidermal involution, but it is evidently 
due entirely to disruption of the central secreting cells whose 
origin has been traced above. 
The opportunities offered by the variety of Nepenthes culti- 
vated at the Royal Botanic Garden, Edinburgh, for observing 
the behaviour of insects have been often taken advantage of by 
me. The warm houses in which they grow are constantly 
infested by colonies of a small brown ant about half the size of 
our fallow-ant. These make nests of from one to three inches 
across in the pots and frame-baskets carrying the plants. On 
a warm day they are constantly running about, and a large 
share of their attention is given to the growing specimens. In 
one of the houses cockroaches are abundant, and the pitchers 
of Nepenthes are favourites with them during night. I may 
now describe an observation made in the summer of 1885. 
Being in the house just indicated at 8.30 p.m. on a clear evening 
in June, a large cockroach was noticed to be perched on the front 
part of the corrugated collar of a fine pitcher of N. khasyana . 
Approaching cautiously, it was seen to bend its head into the 
pitcher-cavity and sweep it rapidly in successive jerks round 
the inner edge of the corrugated collar where the products of 
the marginal glands would lie. Sipping the material from these, 
it then rested for a moment and enjoyed with evident relish 
the cleaning of what adhered to its mandibles. It then 
repeatedly tried with its fore-legs to step on to the conducting 
surface of the pitcher-cavity, but always slipped, so leaving 
this it reared itself on its long hind-legs by planting one on 
each side of the rim, catching with the middle legs on to the 
lower sides of the lid. Placing its fore-legs on the middle 
of the lid, it swept with its mouth-parts the richly honeyed 
surface in long lines. But this did not appear to be satisfying, 
when compared with the product of the marginal glands, for it 
speedily returned to these and renewed its jerking mode of 
feeding. It again attempted to get into the pitcher-cavity, but 
finding this unsafe, it finally licked up traces of the marginal 
gland-secretion which its fore-feet had smeared on the cor- 
