BASIS FOB SPECIFIC CHARACTERIZATION. 11 



accurately specified. No cultures were made. The original masses 

 are assumed to be comprised each of a single species. Parasitism or 

 selective saprophytism is assumed, bul the substratum is rarely 

 designated with sufficient care to make a duplication of the original 

 culture possible. Hence the characters would include whole series 

 of forms whose differences are marked. No account is taken of their 

 omnivorous nature nor of the marked variations introduced in ap- 

 pearance by changed conditions. The method of types and the met hod 

 of substrata as represented by the herbarium material and the pub- 

 lished descriptions we have are. therefore, not sufficient for the 

 identification of species of P< niciUium. 



Alt hough we a hand on the "method of substrata" as the sole has is 

 of description, the force of natural selection as shown by the distri- 

 bution of certain of these species in nature is a most valuable acces- 

 sory. Sonic species occur so constantly upon particular substrata 

 and under particular conditions as to simplify their identification 

 greatly. Examples of this are P. italicum Wehmer and P. digitatum 

 Saccardo, as they are found upon citrus fruits. Unfortunately very 

 few species restrict themselves to particular substrata, so that except 

 for some few species and a very few media, the occurrence of a 

 Penicillium in any given situation is hut slight evidence for its 

 identification. Tin' constant occurrence of species of PenidUium 

 in the laboratory, in connection with foods, and in factory processes, 

 such as cheese-ripening, all point to controlled culture as the proper 

 source of diagnostic characters. 



If we look to cultural study" for our conception of species we have 

 t wo methods of procedure : (1) The exhaustive study of the limits of 

 variability for each species; (2) the comparative study of numerous 

 species under arbitrarily chosen conditions uniformly maintained. 



The first is the best method known for gaining complete knowledge 

 of single species, hut it is too cumbrous for taxonomic purposes. 

 Physiological and chemical studies have commonly been restricted 

 t<> part icula r classes of reactions for single species or groups of species. 

 These have contributed much to our knowledge of fungous varia- 

 bility, but t Hen give no hint, except the vague nomenclature 



used, upon which to judge which species were actually studied. In 

 experimental cultures changes in the chemical nature of the medium 

 or in the conditions, or both, have been found to produce great 

 changes in the morphology of the fungi studied. With the exception 

 of a few fundamental group or generic characters, nearly every 



aWeidemann, ' in a recenl paper, has described several new species of this genus 

 in their relation to culture media. En this he lias followed the bacteriologists fun her 

 than the writer has thought necessary by giving formulas for a considerable number 

 of media and detailed notes a- to reactions upon such media, instead of the more 

 formal descriptions attempted here. 



