272 Reed . — A Study of the Enzyme-secreting Cells in the 
The tissue is allowed to remain in the killing fluid from ten to twenty hours, 
then transferred for washing to 70 per cent, alcohol, which contains about 
1 per cent, of potassium iodide. If the killing fluid is not completely 
removed, it interferes with the action of the stains, more particularly the 
basophil stains. 
Probably the good results obtained with this reagent are due to the 
large amount of soluble chloride it contains, which precipitates the proteids 
in the cells. 
Saturated Solution of Mercuric Bichloride in Absolute Alcohol . This 
reagent was used to kill resting embryos in the dry seeds, where the 
presence of air in the tissues hinders the penetration of heavy liquids. It 
produced good preparations, but did not leave the nuclei of the cells as 
susceptible of staining as Worcester’s fluid. 
It will be seen from the foregoing accounts that the killing fluids which 
gave the best results were those containing a large proportion of soluble 
chloride and a small proportion of acid. Those fluids which contained 
a large proportion of acids appeared to have a corrosive action upon the 
proteid granules in the cells. In small amounts the presence of acids 
appeared to facilitate the penetration of the reagent without corroding 
the granules. 
Most of the material was embedded in paraffin by the ordinary method. 
In the case of dry seeds I used a method for which I am indebted to 
Mr. B. J. Howard (’ 03 ). By the use of this method I completely infiltrated 
the pieces of seeds (which were always small) with paraffin, and succeeded 
in obtaining uniformly good preparations. 
The results obtained by the use of different stains were as variable as 
those of the different killing fluids. No one stain could be depended upon 
in all cases. The use of stains is twofold — to render the objects more 
opaque for study, and to give an indication of their acidity and alkalinity. 
While I did not find it possible to apply the terms ‘ cyanophil ’ (baso- 
phil) and ‘ erythrophil ’ (eosinophil) with precision, yet the absorption 
of different stains serves in a general way to indicate the nature of 
cell-contents. The use of the different stains gave the results described 
below. 
Picro-Nigrosin. This was found to be a very satisfactory stain for 
general cytological purposes. It brings out the granules, nuclei, and cyto- 
plasts very plainly, but does not give them a differential stain. 
Kleinenberg s Haematoxylin . The results obtained by the use of this 
stain were much the same as the preceding. It is a very valuable stain for 
chromatin, and was used chiefly on that account. 
Haidenhain s Iron Alum Haematoxylin. The method described by 
Torrey (’02), using Congo Red as a counter-stain, was employed. It is 
a very valuable stain when used in connexion with others, but cannot be 
