Feb.1,1921 Another Conidial Sclerospora of Philippine Maize 677 



of the length curve of the latter species does not, in the opinion of the 

 writer, indicate that it is bimodal, because, by using more inclusive 

 measurement classes of 4 fx or even 3 fx, the depression so noticeable with 

 the 2-ju classes smooths out and the curve becomes quite regular. More- 

 over, the difference between the modes as well as between the means and 

 the medians is still sufficiently great to emphasize strikingly the dissimi- 

 larity in size of the conidia of the two species. 



It should be noted that, although the curves of frequency distribution 

 of the two species overlap slightly, size is none the less a valuable diag- 

 nostic criterion. In length, for instance, the curves overlap from 26 fx, 

 the lowest limit of the Visayan Sclerospora, to 52 ll, the highest limit 

 reached by Sclerospora philippinensis. As a result, it might be con- 

 tended that size is of no value in distinguishing between the two species 

 when applied at least to the conidia falling between these limits. While 

 this is true of any one conidium, experience shows that, if several are 

 measured, exceedingly few are to be found in this disputed region. For 

 practical purposes even 50 unselected conidia of each species are suffi- 

 cient to show the difference between them without any confusion due to 

 overlapping. 



It is also worthy of note that the curves of the frequency distribution 

 of 700 conidia in both the Visayan species and Sclerospora philippinensis 

 differ in no essential particular from those of 500, 400, or even as few as 

 200 conidia. 



Furthermore, in the ratios of length to width of their conidia, the two 

 species also show marked differences. The shorter, broader spores of 

 Sclerospora philippinensis most commonly show ratios of 1.55 to 2.14, 

 while in 5. spontanea the greater length as well as the lesser width of the 

 coniJia is expressed by the predominant ratios of 2.35 to 2.94. 



In order to determine whether the differences between the biometric 

 characteristics of the two forms were indeed significant, the method 

 quoted by Rosenbaum (11) from Reitz and Smith was employed. This 

 method, which compares the difference between the mean or other con- 

 stants with the probable error of the difference, shows that in Sclerospora 

 philippinensis and 5. spontanea these differences without doubt are 

 significant and can not be the result of mere random sampling. This 

 significance is clearly brought out in Table III. 



TABLE III. — Difference in means of Sclerospora spontanea and Sclerospora philip- 

 pinensis compared to the probable errors 



