September 15, 1905.] 



SCIENCE. 



343 



ning salt water, after a few struggles it 

 turned on its side and so remained, seemingly 

 in considerable distress, being unable to main- 

 tain itself in the normal position by its delicate 

 fHament-like ventral fins which are intra- 

 mandibular in position. 



I then filled a tall glass jar some eight inches 

 deep with fine sand, introduced into it the 

 little fish and placed it under a salt water 

 jet. At first the fish lay quiescent on the 

 sand, but when I returned some hours later, 

 it had burrowed into and was never again seen 

 on top of the sand. Freqiiently, however, the 

 little fish could be seen with its body half out- 

 lined against the glass side of the aquarium. 

 There could then be seen slow undulations of 

 the long dorsal and anal fins together with 

 slight bendings of the body, both motions be- 

 ginning at the head and progressing towards 

 the tail. Evidently by this means a current 

 of water was maintained through the gill- 

 chambers. On the surface of the sand, small 

 conical half-filled depressions could be found. 

 These seemed to have been formed by the fish 

 either in burrowing into the sand or in draw- 

 ing water over the gills. However, I did not 

 notice any distinct currents through these de- 

 pressions and can not positively say that they 

 were excurrent and incurrent openings. But I 

 am sure that there were no distinct burrows, the 

 wet sand not having sufficient consistence to re- 

 main in shape after the withdrawal of the fish. 



Bits of oyster were put into the aquarium 

 as food for the fish, but as these were never 

 counted I could not be sure that any had been 

 eaten and as it was impracticable to make 

 later an examination of the contents of the 

 stomach of the fish, nothing can be said as to 

 its food. 



Since this fish was of no value as a live 

 muse^im specimen, and as I feared that it 

 might die of starvation, it was killed and later 

 identified as Bissola marginata, one of the 

 cuskeels. It is a cause of considerable regret 

 that press of other work prevented a more 

 complete study of the habits of this interest- 

 ing little fish. This specimen is now in the 

 Museum of the laboratory of the United States 

 Bureau of Fisheries, at Beaufort. 



E. W. GUDGER. 



INTERNAL INFECTION OF THE WHEAT GRAIN BY 

 RUST A NEW OBSERVATION. 



There are many species of the rust para- 

 sites (Uredinales) found upon cereals. Al- 

 most every type of cultivated grain is attacked 

 by its own particular type or form of rust. 

 These rusts are minute, thread-like, filament- 

 ous affairs. The threads are very much more 

 minute than the branching spawn met with in 

 mushroom culture. The threads branch very 

 generally and spread through the tissues of 

 the host plant in various directions. They are 

 able to penetrate all of the soft parts of the 

 plants upon which they live. It has usually 

 been supposed that the threads did not spread 

 far from each point of new infection. The 

 filaments usually gain admission to the tissues 

 of the leaves and stems by eroding or boring 

 through the skin layer from the outside. 

 Later the branching filaments become massed 

 at certain points under the skin layer of the 

 host plant. They then produce countless 

 numbers of small ovoid or rounded bodies 

 called spores. These spores are cut off or 

 rounded off from the ends of the filaments, 

 pressing outward under the skin layer of the 

 host. As the spores mature, the size enlarges, 

 and thus the skin or epidermis of the host 

 plant is broken and pushed outward. This 

 allows the spores access to the air and they 

 are then carried by the wind and other agencies 

 from plant to plant and from field to field, 

 perhaps hundreds of miles by wind storms. 

 Countless numbers fall to the ground and do 

 no harm; but countless numbers are produced 

 and thus some of them are sure to reach other 

 host plants. This is the usual method of 

 accounting for the spread of wheat rust. 



It has usually been assumed that rusts grow 

 only in the leaves and stems (vegetative parts^ 

 of their hosts, but gradually it has been 

 learned that amongst many perennials, cer- 

 tain weeds and shrubs these parasites send 

 their filaments (hyplia) into other more per- 

 manent structures, as, for example, roots and 

 woody stems, thus becoming perennial with 

 the host. Observations and experiments at 

 this experiment station have gradually con- 

 vinced us of the probability that rust of wheat 

 may sometimes thus persist. Our field experi- 



