56 H. B. STEINBACH AND PHILIP B. DUNHAM 



Table I. Ionic concentration in Phascolosoma sperm 



The story can be made very short by reference to the data in Ta- 

 ble I. Concentrations noted are approximate and are calculated on a 

 wet weight basis. By our methods, head midpieces account for 75% 

 of the total volume of the sperm and include most of the dry weight. 

 Tails are very dilute in terms of dry weight. Head midpieces, sepa- 

 rated from tails, retain the high ionic gradients of whole sperm. Tails 

 act like so many highly permeable strands of jelly. 



Our results could be explained by assuming that ion gradients are 

 pumped up and maintained by the head-midpiece complex, which 

 leaks ions of its characteristic pattern back through the tail. If this 

 were true, head-midpiece fractions should have higher ion gradients 

 than do whole sperm, since the leaking "parasitic" tails are now no 

 longer present. With Phascolosoma sperm exactly this is found. Head 

 midpieces have more K than do whole sperm; tails have virtually no 

 excess. With Arbacia and Mytilus sperm, the results are less clear-cut. 

 Inulin spaces increase markedly, and ion gradients fall in head-mid- 

 piece fractions. We suggest this is due to injury to some members of 

 the population, the sperm of Phascolosoma being just a harder headed 

 type. 



With regard to mechanisms involved in establishing the gradients, 

 we can offer little at the moment. There is no firm ion binding. Na 

 and K of sperm are freely exchangeable as measured by isotope ex- 

 change with the environment. 



Cellular K is retained tenaciously in the face of marked variation 

 in total ionic strength of the environment with very slight osmotic 

 effects. Likewise, if external KC1 is varied upward from normal sea- 

 water concentration, cellular Na remains remarkably constant. There 

 is no indication of reciprocal changes of internal Na and K as is noted 

 with muscle or nerve. Thus we have no evidence for sodium pumps. 



The usual Donnan ratios for CI and K show no correspondence 



