10 BULLETIN 112, U. S. DEPARTMENT OF AGRICULTURE. 
plants of the grass family and here pass the winter as viviparous 
females; or the winged fail migrants from grain may seek such 
trees as the apple, where the true sexual forms are produced, the 
oviparous females of this generation in turn depositing eggs on the 
twigs and branches, usually in the axils of the dormant buds or in 
crevices in the bark. (Fig. 5.) 
In the latitude of La Fayette, Ind., the species commonly winters 
either as viviparous females on grains and grasses or in the egg 
stage on apple. Farther north, and especially in extremely cold 
winters, this species is probably unable to winter in any but the egg 
stage, while in the southern parts of the United States, where the 
winters are moderate, the aphides may live over winter as vivip- 
arous females only, no egg stage appearing. 
The theory, put forth by Pergande, 1 " that the species is biennial 
and that the progeny of the spring migrants from the apple subsist 
almost exclusively upon various grains and grasses until the fall of 
the second year, when a generation of return migrants makes its 
appearance," is hardly a correct one. The writer's experience shows 
that while the apple may be a fall or spring host of avence, it is not 
a necessary alternate host, and that the species may subsist indefi- 
nitely on grains and grasses, and especially is this probably the rule 
in the Southern States. The species has been reared through more 
than 60 consecutive generations, covering a period of over two years, 
and through three winters on wheat, the warm greenhouse being used 
to carry the species through the winter months, and the line of 
viviparous generations could probably have been continued indefi- 
nitely but for an accident, the aphides having been killed when the 
greenhouse was fumigated without the knowledge of the writer. 
Continuous-generation experiments were conducted at La Fayette, 
Ind., in 1909 by Messrs. W. J. Phillips and T. H. Parks and in 1911 
and 1912 by the writer. In 1909 and 1911 the summers were unusu- 
ally hot, and the experiments were not satisfactory, but in 1912 it 
was possible to get continuous first-born and last-born generation 
series without breaks. In 1909 Phillips and Parks obtained a maxi- 
mum of 15 generations from May 15 to October 7 and a minimum of 
8 generations in the same length of time. In 1911 a maximum of 
18 generations was obtained from April 29 to October 12, and in 1912 
a maximum of 23 and a minimum of 9 generations from May 3 to 
November 13, or a mean average of 16 generations. In the Southern 
States, where the species may breed throughout the winter months, a 
much greater number of generations would occur. In the experi- 
ments of 1909 the average number of young per female, in the 21 
cages where records were kept, was 30.6; in 1911 the average for 
1 U. S. Dept. Agr., Div. Ent., Bui. 44, 1904, p. 7. 
