io6 
Journal of Agricultural Research 
VcJL XXV, No. 2 
In fields B and D, which were rogued of diseased hills as soon as the 
symptoms were observed, the small amount of infection found by August 
12 may be the result of insect transmission occurring before the roguing 
was performed. A comparison of Series 3, 10, and 11 shows that as much 
mosaic may be contracted by plants in a healthy plot in the same field 
with mosaic plants and plots as by plants grown on soil that supported 
a mosaic plot the year before (40 , p . 335-36). A comparison of Series 
1 and 6 shows that considerable mosaic infection reached the tubers of 
hills next to diseased hills by August 13 in one field, but not in the other, 
and a comparison of Series 2 and 8 shows that it was only in the former 
field that leaf-roll infection reached the tubers of hills next to diseased 
hills by September 15. The former field was about three times as far as 
the latter from the nearest known winter host of the potato aphids (36) 
and possibly became infested later or to a less extent, or both. 
In 1921 hill selections were made in two Irish Cobbler fields contain¬ 
ing, respectively, 5 and 15 per cent of hills producing spindling tubers, 
and in a Green Mountain field containing 1 per cent. The results, given 
in Table III, indicate that hill selection was useless for eliminating dis¬ 
ease, though it reduced the increase otherwise occurring without hill 
selection. It is not surprising that hill selection in a field containing 
disease is often disappointing. Murphy (29, p. 45-47) and Quanjer (39, 
p. 142) have had the same experience. When as low as 5 per cent of 
diseased hills may contaminate the majority of the healthy hills in the 
same field, the hill-selection method clearly has limitations. 
REMOVAL OF DISEASED HILLS 
The removal of diseased hills, or roguing, is in the broad sense a 
method of hill selection. It differs from hill selection in the strict sense 
chiefly by using healthy hills after removing the diseased hills that may 
serve as sources of infection, whereas the latter uses healthy hills that 
have been more or less exposed to infection throughout the season. The 
effectiveness of roguing is determined by several factors. Correct 
diagnosis, thoroughness at each inspection, complete removal of each 
rogued hill from the field, and several inspections at the proper times 
are necessary. 
Roguing with no insect spraying and with no great degree of isolation 
has been tested for several years, 1917-20, by the writers (45 , p. 270; 
4°> P- 332; 41 , p. 77). With variations according to the season and 
degree of proximity to diseased plots, the general results have been to 
keep the amount of mosaic in the stock about the same from year to 
year, between 13 and 30 per cent, except that roguing in 1920 was 
followed in 1921 by 35 per cent and 65 per cent in Green Mountains and 
Bliss Triumphs. This was a gain over conditions in unrogued parts of 
the same strains. Why the removal presumably of all diseased hills 
has not eliminated the disease, or at least steadily decreased the per¬ 
centage, is not certain. Possible causes are masking of symptoms at 
the time of roguing, dispersal of transmitting insects from rogued plants 
before or after these plants are pulled and insect dispersal from unrogued 
plots. 
Roguing accompanied by isolation from unrogued diseased stock and 
by insect spraying will be considered later. 
