13 



on the surface of the water and tlie wliole can be phiced on the 

 table of the microscope. If we renew the water now and then, 

 and if, after examination, we put the prcparation back into 

 tlie aquarium, we can observe this living- tissue under wliolly 

 normal conditions for many hours, on several consecutive days. 



Microscopising . All my microscopic examinations were made in 

 an Engelmann case, with the help of Zeiss's eye-piece n°, 4 

 and a most excellent specimen of a Leitz '/,.^ oil-immersion. This 

 last lens gave even better results than the precieus apochromatic 

 system of Zei.ss (with compens. eye-pie?es). 



The estimating of the nuinher of (•hlorophijll rorpusdes^ etc. in 

 a preparation. I often had to know — for the sake of mutual 

 comparison — the absolute number of, for instance, the various 

 green and colourless chlorophyll corpuscles in a same volume of 

 several sponges, or in a certain volume of several cultures. If 

 I had wished to count these corpuscles, it would have required 

 a suspension — adequate to counting as for exactness — of equal 

 volumes of sponge-tissue (or culture) in the same quantities of water; 

 otherwise the exact method of counting would have been worth- 

 less. But of course, we cannot obtain equal volumes of sponge- 

 tissue or culture, nor a really equal suspension ; while, also, 

 counting would have proved a very tiresome work. 



I therefore always took, by means of pincers from a sponge 

 or with a pipette from a culture, quantities of the material as 

 equal as possible, and spread it out on a glassslide, while parts 

 of the skeleton were removed. Then the coverglass was pressed 

 in an always equally stroug way, the superfluous liquid sucked 

 up, and the coverglass surrounded by vaseline ; next I estimated 

 under oil-immersion the number of the various chlorophyll cor- 

 puscles present in the whole microscopic preparation, consequently, 

 present in an almost equal volume of each sponge or of each 

 culture. This method, always applied in the same manner, gave 

 rather good results, as Table n°. 4, 6, 8 show; besides, the dif- 

 ferences, I wanted to know, were generally considerable enough. 

 So this method of estimating was in all respects preferable to 

 counting, which would have led to an imaginary exactness only. 



