555 



vegetation, in which green flagellates and small algae and dia- 

 toms are most abundant numerically, and, therefore, quanti- 

 tatively. Moreover, in 46 of the 105 instances the movement 

 in production, as shown in the rise and fall of the plankton 

 taken by the filter-paper and silk net, is in the same diredioti, 

 though amplitudes reached are rarely proportionate. The co- 

 incidence is most marked when the catches of the silk net re- 

 veal changes in production of considerable magnitude, as, for 

 example, during the rise and fall of the vernal pulse in 1897 

 and in 1898, and during the winter changes of 1898-99. There 

 are suggestions in these records of vernal pulses of considerable 

 magnitude, of a large midsummer production, and of a great 

 development in the low water of 1897, when an enormous 

 growth of Chlami/douionas turned the river to a livid green, and 

 contributed to the maximum filter-paper record of 119 cm.^ per 

 m.^ 14-fold the coincident catch (8.47) of the silk net. There 

 is also, even in these erratic data of the filter-paper catches, 

 some evidence of the pulse-like character of the production of 

 these minute organisms which form the greater part of the 

 catch. This appears often to be coincident with the cyclic 

 movement in the volumetric data of the silk net, and may best 

 be seen in the records of the winter of 1898-99. The enumer- 

 ation confirms beyond all question the existence of these recur- 

 rent pulses, dimly suggested in these volumetric records. The 

 spring and summer plankton which leaks through the silk is 

 largely made up of small algge, flagellates, and diatoms, with 

 some ciliates, principally Codonella, and rhizopods. The winter 

 plankton thus lost is largely composed of broken colonies of 

 Si/uura, together with many predatory and elusive Infusoria, 

 largely representatives of the Holofricha, which multiply abun- 

 dantly with the autumnal increase in bacteria. 



It can be apparent to no one more than to the writer that 

 such data as these are unsatisfactory in determining the pre- 

 cise volumetric extent of the leakage through the silk. That 

 this leakage is, however, beyond all question considerable must 

 become evident to any one who actually works over collections 



