154 



are also distinct." (See PL IX, No. 20.) Conidial length is given in 

 Fig. U, Graph 100; breadth, in Fig. V, Graph 102; septation, in Fig. T, 

 Graph 88. (See also photomicrographs of conidia in PI. XXVI, d.) 

 By comparing these graphs and data with those of H. No. 3 and H. No. 1 

 it will be seen that the conidia are quite short and a triflethick. The 

 coefficient of cylindricity 78 for subcylindrical conidia, 74 for the more 

 elliptical ones is the highest coefficient shown in any of my strains (cf. 

 with page 119). It seems very probable, therefore, that either the culture 

 was contaminated in Dr. Westerdijk's laboratory or the other culture in 

 Dr. Bakke's hands; or that Dr. Bakke's description was erroneous or that 

 one of the cultures was contaminated by me; or that the facts represent 

 a real hereditary change in morphology during a long series of transfers 

 and I incline strongly to the last of these alternatives. 



H. No. 23 showed what appears to be a modification rather than a 

 saltation, in that in the original culture received from Miss Weniger there 

 were many abnormal tri-pointed conidia (see page 101). It is suggested 

 that a change somewhat like this, if permanent, may have given rise to 

 the forms with unequal central cell, for example, to H. No. 29. 



All of the above-mentioned examples appear to represent clear-cut 

 cases of change in morphological character in test-tube culture. The 

 only essential difference between these changes in test-tube culture and 

 the saltation reported in Petri dishes is that in the cases of the Petri dish 

 selection of the saltant was voluntary, while in making transfers from 

 tube to tube the selection was accidental. 



SALTATIONS IN NATURE 



That changes do occur under my culture conditions renders it highly 

 probable that they also occur in nature in the fields. Thus a saltant 

 strain may become established on one wheat plant, form large numbers 

 of conidia, gain foothold in a region, and then enlarge this foothold, per- 

 haps to cover large areas. That one strain may thus outgrow another 

 has been shown by Crabill (36) and is evident in my own work (PI. XXX, 

 XXXI, lower figs.). The fact that so many strains of Helminthosporium 

 differing slightly but distinctly from each other, yet agreeing closely in gen- 

 eral, can readily be isolated from cereals, indicates that probably this also 

 has naturally happened, and that in the fields we have today large numbers 

 of races or strains of closely related forms derived more or less recently from 

 a common parent stock. To test this hypothesis experimentally, in field or 

 greenhouse, by inoculation with pure cultures, and later isolating organ- 



