208 



CARNEGIE INSTITUTION OF WASHINGTON. 



It is necessary to add the alkali as bicarbonate, in order to avoid precipi- 

 tation of MgCOs and CaCOg, but after the mixture is made air must be 

 bubbled through it rapidly for 6 hours, and longer if the rate is slow, in order 

 to remove part of the CO2 resulting from the dissociation of the bicarbonate, 

 and thus obtain the correct hydrogen-ion concentration. In table 6 is given 

 the approximate concentration of some minor constituents of natural sea- 

 water and their probable concentration in artificial sea-water made from 

 "analyzed" or "chemically pure" salts in parts per million. This estimate 

 is confirmed by the fact that further recrystallization of the salts in pyrex 

 glass did not improve the artificial sea-water. Artificial sea-water made 

 according to this fomiula was satisfactory for animals. Plants require PO4 

 and NO3 or NHs, but 1 NH3 or 250 NO3 per million is toxic to some algae 

 (but not to some diatoms) , whereas animals with symbiotic alga? were affected 

 by 5 PO4 per million. 



Table G. 



In previous papers I have shown that certain ions of sea-water increase 

 the permeability of cells and that this change is inhibited by the other ions. 

 This year's experiments on Cassiopea and the heart of the conch indicate 

 that 0H~, Na+, and K+ increase the permeability and that H+, Ca++, 

 and Mg++ inhibit their action. It should therefore be possible to find a 

 protective solution for one cell structure containing only two of these ions 

 in a balanced ratio. The fact that attempts to do this have failed indicates 

 that each cell (of those thoroughly investigated) is composed of more than 

 one structure with a characteristic optimum ratio for each pair of antago- 

 nistic ions. It is important, therefore, to intensively study very special 

 phenomena, as Dr. Mayer has done in regard to nerve-conduction, in order 

 to collect data for an analysis of the phenomenon of antagonism. 



It has long been known that a nerve of the conventional nerve-muscle 

 preparation may be stimulated electrically by less current if it is sent longi- 

 tudinally through the nerve than if it is sent crosswise; but the explanations 

 of this fact have not been in accord with recent work on the phenomenon 

 of stimulation. Since the nerves that have been used were medullated, 

 I repeated this experiment on the non-medullated nerves of Limulus and 

 found that they were stimulated crosswise by as little current as was found 

 to be the minimum for stimulation lengthwise (the electrodes being the same 

 distance apart in each case). Therefore, the medullary sheath seems to 

 inhil)it cross-stimulation. Probably the medullary sheath is an insulator. 

 The electrical conductivity of nerve is greater lengthwise then crosswise. 

 Perhaps the alternating current running lengthwise between the nerve-fibers 

 induces currents inside the medullary sheaths; or the current may enter and 

 leave only at nodes of Ranvier and the poor result of cross-stimulation may 

 be due to the small number of these nodes between the electrodes, unless 

 the length of nerve affected is increased by spreading of the current (as is 

 the case when stronger currents are used). 



