Methods of Investigation. 7 



protectors, but afterwards I found covers made of gauze 

 and provided with a glass top were more suitable : they 

 can easily be made of any size, are convenient for 

 observation, and allow of free ventilation ; besides 

 under them a tree can be watched for days together, 

 whereas a bell-glass protector soon becomes dimmed 

 with moisture and requires frequent cleaning. 



The oak saplings I have sometimes grown and some- 

 times obtained from nurserymen. The four to six year 

 old saplings are to my mind the most convenient size, 

 and a large choice of them makes experimental breeding 

 much more easy. I employed almost entirely Quercus 

 sessilijlora. It is essential that a sapling about to be 

 used in an experiment should have its buds well 

 developed, as these are always preferred by the flies. 

 There is a difficulty in rearing species which only 

 prick flower-buds, since young saplings which do not 

 produce flowering catkins are useless for the purpose. 

 The only way to rear those species is to make the 

 experiments on full-grown trees in the open air, 

 taking every means to guard against possible error \ 

 On the other hand it is very easy to make experiments 

 on saplings with species which prick the leaves or 

 bark. Such briefly was the plan adopted by me in 

 investigating the development of the species which 

 I am about to describe. 



As it is very difficult, and in some cases almost 

 impossible, to distinguish the flies of nearly related 

 species from each other, any illustrations of the insects 



\} Dr. Beyerinck used for this purpose cubes of wire covered with 

 muslin and tied round the branch.] 



