154 



SCIENCE 



[N. S. Vol. XLVII. No. 1207 



mark one pole and the neurite tlie other. 

 This idea of the polarity of the neurone 

 seems to have originated with van Gehuch- 

 ten (1891), but on its announcement it was 

 immediately taken up, amplified and 

 strongly advocated by many other neurolo- 

 gists, especially by Ramon-y-Cajal (1891) 

 and by Retzius (1892) . Even now it holds 

 a place in good text-books such as Herrick's 

 Introduction to Neurology (1915), but it 

 has never been without its serious oppo- 

 nents. 



Barker (1899) in his account of the nerv- 

 ous system pointed out numerous and seri- 

 ous exceptions to it. The intermediate neu- 

 rones of the vertebrate retina (Fig. 2, i) 



Fig. 2. Diagram of the chief nervous elements 

 in the vertebrate retina; rod- and eone-ceUs (r), 

 intermediate cells (i), and ganglion cells (g). 



receive impulses by one system of processes 

 and discharge them by another ; yet the dis- 

 charging processes have none of the struc- 

 tural peculiarities of neurites, but closely 

 resemble the receiving processes, which are 

 indistinguisha;ble from dendrites. Again 

 the sensory neurones of the spinal and the 

 cranial nerves (Fig. 1, d) have their cell 

 bodies on their length and not terminal, 

 but their distal processes, which by defini- 



tion should be dendrites, are from the 

 standpoint of structure in all respects neu- 

 rites, as are their proximal processes. 

 Hence arises the difficulty that in some in- 

 stances dendrites, though usually receptive 

 in function, may discharge, and neurites 

 ithough usually discharging in function, 

 may receive. These and other like causes 

 make it clear that any attempt to define 

 neuronic polarity on the basis of dendrites 

 ,and neurites is foredoomed to failure, for 

 (Whatever may be the explanation of the 

 idifference between the two classes of proc- 

 esses, it is not necessarily connected with 

 the direction in which they transmit im- 

 pulses. This conclusion, moreover, is sup- 

 ported by what is known of invertebrates. 

 pSere it is commonly quite impossible by 

 ,any of the structural tests used in verte- 

 brates to distinguish dendrites from neu- 

 rites. In fact the processes at the two poles 

 of the cell seem to be essentially similar and 

 both tend to resemble dendrites. 



But in my opinion the chief reason the 

 hypothesis of neuronic polarity as ordi- 

 narily stated meets with serious difficulties 

 is not because of the complications that 

 arise when dendrites and neurites are in- 

 volved, but because nervous activity is 

 commonly described in reference to the 

 cell body of the neurone. Certain parts of 

 the neurone are said to transmit toward 

 the cell body or cellulipitaUy and other 

 parts are said to conduct away from this 

 body or cellulifugally. But polarity es- 

 tablished upon this basis is surely upon 

 very insecure ground, for it involves the 

 assumption that the cell body of the neu- 

 rone is the center of nervous activity, an 

 assumption which has come down to us 

 from the past, but in support of which lit- 

 tle can really be said. 



To be sure neurofibrils have been abun- 

 dantly identified in the cell body of the 

 neurone, but they have never been shown 



