116 PROCEEDINGS OF THE ACADEMY OF [1898. 



blue, fuchsin, etc., for the osmic acid seemed to produce, as it were, 

 a certain shrinking of the very delicate organs. A measure of suc- 

 cess followed the use of gentian violet followed by tannic acid, both 

 in minute quantity. By this means the coleoderm was outlined; 

 and especially in the case of frnstules that had passed the period of 

 their activity, and were nearly or quite dead, as shown by their con- 

 tracted nucleus and bacteria-infested outline, the surface staining of 

 the gelatinous external layer was clear. An extreme case of this is 

 roughly shown in fig. 7, PI. VI. Here the contents of the inner cell 

 were apparently normal, with the exception of the nucleus, but the 

 coleoderm was quite flaccid, loosely adherent to the frustule, and 

 abnormally enlarged, as well as fringed with colonies of bacteria. 



The stain which most quickly and surely shows the healthy coleo- 

 derm, at the same time instantly killing the diatom, is made as fol- 

 lows: — - 5 gram of Bismarck brown and 1.0 gram tannic acid are 

 dissolved separately and added to a liter of distilled water. The 

 solution remains perfectly clear, and is of a reddish-brown color. 

 Two or three drops of this are added to a drop of water, under the 

 cover glass, containing errant frustules of Eunotia. Almost before 

 a change of tint is visible in the thin layer of water under the mi- 

 croscope, motion ceases, and, at the same time, at each corner of 

 the frustules appears a little rounded mass of substance, gelatinous 

 in appearance, and dotted with coppery or bronzy specks of most 

 minute size. The shape of these masses, and their relation to the 

 frustule, are indicated in figures 4 and 5, PI. VI. The results of 

 somewhat heavier staining with the same mixture are shown photo- 

 graphically by Mr. F. J. Keeley in figures 4, 5, and 6, PI. VII. 



Pursuant to the further study of these coleoderm masses, a large 

 number of errant frustules were stained as follows : About two drops 

 of water, containing the diatoms, were put on a cover glass and 

 allowed some minutes to bring themselves into normal relations with 

 the glass. Three drops of the described stain solution were added 

 and staining was allowed to proceed half an hour. The excess of 

 color was then removed by careful washing, dipping the cover re- 

 peatedly into a cup of water with as little friction as possible. 

 Griibler's aqueous eosin, diluted with an equal quantity of water, 

 was now added and allowed to act half an hour to an hour. Finally, 

 the glass was washed and mounted in very weak formalin. The di- 

 atoms, under these circumstances, remain attached to the cover glass, 

 as a rule with ventral side uppermost, and with coleoderm processes 

 in the position best suited for study. 



