274 



of its excreta. A detailed account is given of experiments extending 

 over four months which establish the fact that P. longipes has a 

 pecuharly favourable influence on C. viridis ; on ant-infested bushes, 

 the death rate of the scales is considerably lower, they develop more 

 rapidly, their parasitation by Ichneumonids is reduced, and their 

 progeny is twenty times more numerous. The suggested explanation 

 of these facts is that feeding is promoted in scale-insects in which 

 excretion is accelerated by the stroking performed by P. longipes. 

 It was noticed that feeding seems greatly to influence the development 

 of the scale ; on the same coffee bush the young scales on the shoots 

 develop more rapidly than those on the leaves, which form a less 

 suitable food. 



In investigating whether other ants also benefit Coccus viridis, the 

 choice naturally fell on the black cacao ant DolicJwderus, hiliiberculatus, 

 which protects cacao against HelopeUis [see this Revietv, Ser. A, v, 

 p. 143]. D. bituberculatvs is seldom seen in plantations of Coffea robusta, 

 but was frequently met with in those of Liberian coffee and, when this 

 was replaced by cacao, the latter became infested. Expenments 

 similar to those with the gramang ant showed that the death rate 

 of the first generation of the scale was practically the same whether 

 D. bituberculatus was present or absent ; the development of C. viridis 

 occupied on an average 83 days when no ants were present, 

 74 in the presence of the black cacao ant, and only 65 in that of the 

 gramang ant ; the average percentage of Ichneumonid parasitation 

 was 39 per cent, when no ants were present, 3*5 per cent, in the 

 presence of the black cacao ant, and 5 "7 per cent, in that of the 

 gramang ant ; the average number of scales after 4|^ months was 

 70 on ant-free bushes, 403 on bushes infested with black cacao ants, 

 and 1,057 on those infested with gramang ants. It is suggested 

 that this difference in the influence of the two species may be due 

 to differences in their behaviour. Only a few individuals from a 

 gramang colony perform the milking ; they move quickly over the 

 scales and stroke lightly and for a short time only. On the other 

 hand many members of a black cacao ant colony are employed and 

 stroking is both heavy and prolonged, so that the sensitiveness of 

 the scales is blunted instead of being stimulated. 



Other ants occurring on Javanese coffee estates include Oecophylla 

 smaragdina, Pheidologcton diversus, Cremastogaster spp., Camponotus 

 camelinus, Myrmicaria brimnea, etc. Though no experiments were 

 made, it was noticed that the advent of various ants among ant-free 

 colonies of C. viridis always led to an increase of the latter. Some 

 planters have attempted to drive away P. longipes by introducing 

 other species, but, as C. viridis invariably thrives better when ants 

 are present, not only P. longipes, but all other species should be 

 kept away from coffee estates as much as possible. 



The white cacao scale, Pseudococcus crotonis, Green, is abundant 

 on, but apparently harmless to the pods and twigs of cacao, 

 in the plantations of which it forms the chief attraction for 

 D. bitubercidatus. Under experimental conditions it was ascertained 

 that P. crotonis died out on cacao trees uninfested by ants, while it 

 flourished on ant-infested ones. Under natural conditions the scale 

 usually disappeared, when the gramang ant was present, because this 

 ant does not sufficiently protect it (as the black cacao ant does) 



