76 METHODS IN PLANT HISTOLOGY 



Probably the most important reason which led Fischer to undertake 

 this series of experiments was the claim that certain granules of the 

 Cyanophyceae should be identified as chromatin because they behaved 

 like chromatin when stained with haematoxylin. Fischer's experi- 

 ments not only proved that chromatin cannot be identified in this way 

 but raised the question whether staining reactions ever indicate chem- 

 ical composition. At present, it would seem that, in most cases, the 

 staining indicates only physical differences. However, in some cases 

 there is a chemical reaction, e.g., when material fixed in bichloride of 

 mercury is stained in carmine, mercuric carminate is formed. 



It would be very convenient if we knew just how much dependence 

 should be placed upon staining reactions as a means of analysis. If 

 two structures stain alike with Delafield's haematoxylin, does this mean 

 that they have the same chemical composition; or if, on the other 

 hand, they stain differently, must they necessarily be different in their 

 chemical composition? Delafield's haematoxylin, when carefully used, 

 gives a rich purple color, but a careful examination will often show 

 that in the same preparation some structures stain purple, while 

 others stain red. Does this mean that the purple and red structures 

 must have a different chemical composition? Many people believe 

 that structures which stain differently with a given stain must be 

 chemically different, but they readily agree that structures which 

 stain alike are not necessarily similar in chemical composition. 

 Chromosomes of dividing nuclei and lignified cell walls stain ahke with 

 safranin; chromosomes and cellulose cell walls stain much alike with 

 Delafield's haematoxylin; but everyone recognizes that the chromo- 

 some is very different in its chemical composition from either the cellu- 

 lose or the lignified wall. 



However, in an indirect and somewhat uncertain way, one can in- 

 fer the nature of certain structures from the staining. For instance, if 

 sections of various objects have been stained with safranin, we may 

 draw the following inferences with more or less confidence : if cells in 

 the xylem region of a vascular bundle stain red, their walls are ligni- 

 fied; if cortical cells, which may appear quite similar in transverse sec- 

 tion, stain red, they are likely to be suberized; if the outer walls of 

 epidermal cells stain red, they are cutinized; but if the outer boundary 

 of the embryo sac of a gymnosperm stains red, it is chitinized. Of 

 course, these inferences can be made only because the various struc- 

 tures have been tested bv more accurate methods. 



