INTKRPRETATTON OF PARASTTE STATISTICS. 85 



8. Certain parasites with more or less established habits require 

 that their hosts be in certain habitual locations (for example, Neocato- 

 laccus requires stem-dwelling hosts), and in like manner there are 

 conditions which can be made more favorable for parasite attack 

 through cultivation or through plant selection. Furthermore, since 

 parasites require different conditions, it is desirable to alter the 

 existing conditions so as to make them favorable for as many species 

 as possible. In the case of insects extending their range over many 

 different climates it should be the aim to introduce parasites best 

 adapted to the prevalent conditions. 



2. INTERPEETATION OF PARASITE STATISTICS. 



From the great mass of parasite statistics given in Part I a number 

 of important facts need to be consideredo 



Parasitic and predatory attack is strongest from August until frost 

 time. Hence it may be presumed that whatever artificial propaga- 

 tion is to be done will be most profitable when conducted during this 

 period, provided it does not interfere with early fall destruction of 

 stalks, the fundamental cultural remedy against the boll weevil. 



The greatest control of the boll weevil by insects and also by 

 all agencies is in hanging squares. As has been stated in Part I 

 (sec. 3), the hanging squares are a result of a diagonal absciss layer, 

 which causes the drying square to fail in separating itself completely 

 from the plant. These squares die on the plant and aft'ord a very 

 favorable position for parasitic attacks upon the weevils within. 

 The statistics show that insect control in fallen squares is greatest in 

 the moist States of Louisiana and Mississippi. This is undoubtedly 

 due to some of the new parasites which are accustomed to attacking 

 woodland weevils and other insects characteristic of this humid 

 region. The insect control in hanging squares is the greatest in the 

 comparatively dry States of Texas and Oklahoma. These dry States 

 also have a higher combined natural control in the fallen squares 

 than in the hanging squares, largely because the climatic conditions 

 cause a higher mortality of weevil stages in squares lying on the 

 heated surface of the ground. On the contrary, the humid States 

 have a higher mortality from both climatic and insect agencies in 

 hanging squares than in fallen. Furthermore, it has been proven, in 

 Part I (sec. 4), that an increase in the amount of hanging squares 

 will increase the total control. Having these facts in mind, the 

 obvious conclusion is that it will be desirable to have varieties of 

 cotton which have this tendency best developed. Among the varie- 

 ties which are now known to retain their squares are the cluster 

 varieties, including the Rublee. 



