121 
larly in a—e, b—e, c—e, d—e, and e—e’ is significant. If plate e be left 
out of consideration, the others agree reasonably well, with differences 
greater than the probable error in six out of ten cases, the difference being 
but slightly above the probable error in two cases, about twice the probable 
error in two cases; and about thrice that, in two cases, the largest excess, 
in plates a—b, being 0.70.21. 
Plate e deviates widely, with a difference in case of a—e of 2.03 +24, 
the difference being more than eight times the probable error. The great 
difference in plate e must indicate variability of the fungus on this plate 
(cf. with page152),modification due to influence of some unknown factor of 
environment, or error in sampling. But since such a variation did occur 
in a series of plates made with the greatest care and with the same organ- 
ism, it is clear that the occurrence of such a differenee can not properly be 
interpreted as meaning specific difference. Data from the combined rec- 
ords of a, b, c, d, and e’ (omitting e as questionable) give the most reliable 
data I have on length of conidia of H. No. 1 under standard conditions 
(cf. with Graph 42, Fig. K). 
To determine how wide a variability occurs in specimens collected in the 
open, on the natural host, H. ravenelii (P|. XX) a well-marked, easily recog- 
nized species of wide geographic distribution, growing on Sporobolus, was stud- 
ied in conidial-length graphs made from specimens listed in connection with 
the graphs (Fig. L). The tabulated results of this study of H. ravenelii follow: 
Nos.* Differences between means Nos.* Differences between means 
43—46 0.48 + .26 46—50 0.50 + .28 
43—47 0.59 + .24 46—51 0.70 + .28 
43—49 0.62 + .28 46—52 1.46 + .28 
43—48 0.62 + .29 46—53 2.43 + .27 
43—50 0.98 + .27 47—51 0.59 + .26 
43—51 1.18 + .27 47—52 1.35 + .26 
43—52 1.94 + .27 47—53 2.33 += .25 
43—53 2.92 + .26 48—S1 0.56 + .30 
44—48 0.45 + .31 48—52 1.32 + .30 
44—49 0.46 + .31 48—53 2.30 + .29 
44—50 0.81 + .30 t 49—51 0.55 + .29 
44—51 1.01 + .30 49—52 1.31 + .29 
44—52 1.77 + .30 49—53 2.29 + .29 
44—53 2.75 = .29 50—52 0.96 + .29 
45—50 0.80 += .24 Si—52 0.76 + .29 
45—51 1.00 + .25 D153 1.73 + .28 
45—52 1.76 + .25 52—53 0.97 + .28 
45—53 2.74 + .24 
*For significance of numbers, see Figure L. 
