CACTUS-FEEDING INSECTS AND MITES 79 



In the case of prickly pears the attack is confined to the flowers and 

 fruit. Of the Cylindropuntia O.fulgida both the fruit and the very young 

 soft fleshy growth are eaten; sometimes the new growth is preferred. 

 Usually the larvae infest the flower buds, entering at the apex of the 

 petals. They feed within the buds and in the open flowers, devouring 

 the stamens, style, and fleshy portions of the petals; later they tunnel 

 down into the ovary and eat out the seed cavity. Development is often 

 completed within the flower head. When fruit are attacked, entrance 

 is generally through the crown, which is covered with reddish webbing 

 containing frass pellets; more rarely the fruit is penetrated from near 

 the base. The larvae are active, crawling freely over the plant to 

 transfer from fruit to fruit. Frequently two or three larvae inhabit one 

 flower head, and as many as eight have been counted in an unopened 

 bud. As regards injury to the plant, the feeding in the buds and flowers 

 prevents seed development, since the fruit are rendered sterile; and 

 when the fruit is entered, the seed cavity is generally eaten out. 



The duration of the larval life has not been ascertained. The active 

 larval stage is comparatively short, probably occupying two to three 

 weeks, and in any case less than a month. The full-grown larvae descend 

 to the ground and form small circular cocoons of fine strong silk 

 covered with earth. Cocoons have not been located in the field, but in 

 cages were attached to the woodwork below the soil level. Pupation 

 does not take place for some time; for example, larvae that formed 

 cocoon cells at the end of April had not pupated by the end of August; 

 probably they remain in this condition for the greater part of the year. 



N. elautalis appears in the spring and early sum.mer concurrently 

 with the flowering of prickly-pear. In Texas the larvae occur in April 

 and May, the latest date for full-grown larvae in the field being May 19. 

 Larvae were in various stages of development in Valverde County 

 on April 19, 1925; a week later many were full-grown, and some 

 had already vacated the fruit; by May 9 a few late individuals re- 

 mained. In New Mexico the main occurrence of larvae is from the 

 latter part of May to the third week of June. 



Observations on the life cycle are incomplete, but the various data 

 indicate that there is one annual generation, with a second or partial 

 second generation occurring in southern Arizona. Thus larvae col- 

 lected in Texas in April were still unpupated larvae in cocoons at 

 the end of August, while in the field larvae were found in this state 

 in April and May only. 



As regards the second generation in southern Arizona, the following 

 records are offered. Around Tucson in 1924 the main attack appeared 

 to have ended some time prior to the third week of July, when a few 

 larvae, presumably of a second generation, were noticed in the fruit 



