123 



paper and the conidia allowed to develop to maturity. Conidial length 

 here (Graph 64, Fig. O) was somewhat less than under standard conditions 

 (see Graph 42, Fig. K), and the coefficient of variability was a little high. 



Conidial breadth.- H. No. 1 was quite constant in conidial breadth 

 as follows: 



M (7 CV 



6.03 .04 0.55 .34 9. 13 .57 



The ratio of conidial length to conidial breadth is an important one 

 as determinative of shape. This ratio for H. No. 1 is as follows: 



mean length 22.62 .05 _ 



o / 4 ^ . Uo 



mean breadth 6.03 .04 



In a description of H. No. 1, written in May, 1919, for my own use, 

 and prepared with considerably more care than is ordinarily used in specific 

 descriptions of fungi, I noted the conidia as 3 8 septate and as 52.6 67.2 

 XI 9. 2 24 IJL long on wheat; and as 48 84X18 21.6 /JL on corn-meal 

 agar, whereas my more extended study now shows the mode on wheat as 

 78.2 /z, the mean as 76.8 M and the range from 34 to 98.6 /r, the breadth as 

 ranging from 17 to 23.8 /-i, with the mean as 20.4 /*; the septa with a 

 mode of 8, a mean of 7.9, and ranging from 4 to 10. I may here note also 

 that Bakke (6) in his description of H. teres gives the conidial dimensions as 

 150 [or 105*] 130X15 20 ju, and the septa as 7 14. Thus he seems to have 

 found conidia considerably longer than I did, as also narrower ones. It 

 should be said that the data obtained by this study of graphs of H. No. 1, 

 though involving several thousand measurements, fail to record the longest 

 conidium observed, and the one with the most septa, because these were 

 both seen during observations which rendered their inclusion impossible; 

 which is to say that to include them would have been to consciously 

 select these unique conidia for inclusion. Anent the shortcoming of my 

 own brief description cited above may be quoted the Saccardian de- 

 scription of H. ravenelii: "Spongiosum; hyphis flaccidis flexuosis nodosis 

 ramosis, inarticulatis; conidiis cymbiformibus, 3-4 septatis, fuscis, 50 

 ju longis, endo-chromotibus isthmo connexis." Though the mode is approx- 

 imately at 50-54 IJL the conidia really range from 13 to 7lju (see Fig. L). 

 Very similar errors, due to brevity of description, exist regarding many 

 or all known species. 



*See Pammel, King, and Bakke (90, p. 180). 



