September 30, 1904.] 



SCIENCE. 



437 



reaches about latitude 54° and there divides 

 into a northerly and southerly stream. At 

 this point the maximum temperature of the 

 water is 68° F., but the average about 60° F. 

 As it moves doven the coast it loses its heat 

 and produces the rains- and fogs of the 

 Oregonian region, cooling off so that when it 

 reaches the latitude of the Golden Gate it has 

 only the temperature of 54°, or thereabouts, 

 and is colder than the normal sea water for 

 that latitude. It continues southward as a 

 cold current as described by Dr. Bishop. 

 That this essentially superficial stream is not 

 due directly to the impinging of cold Ant- 

 arctic water on the northwest coast seems to 

 be certain from the fact that the temperature 

 of the latter is only 32° F., while the current, 

 when it first reaches the coast, is thirty de- 

 grees or more warmer than that; and also 

 that the water of the current is warmer in 

 latitude 54° than it is in the more southern 

 part of its course, whereas, if it was abyssal 

 water we should expect it to be colder and to 

 gradually warm up as it moved sovithward ex- 

 posed to the action of the sun. 



Wm. H. Dall. 

 Smithsonian Institution, 

 Washington, T>. C, 

 September 12, 1904. 



SPECIAL ARTICLES. 



ON THE PUPATION OF ANTS AND THE FEASIBILITY 



OF ESTABLISHING THE GUATEMALAN KELEP 



OR COTTON-WEEVIL ANT IN THE 



UNITED STATES. 



In a recent number of Science Dr. 0. F. 

 Cook publishes some of Mr. G. P. Goll's obser- 

 vations on the cocoon-formation of the kelep 

 ant {Eciatomma hiberculatum Oliv.), appar- 

 ently in the belief that similar observations 

 have not been made on ants with enclosed 

 pupae. As a matter of fact, it has been known 

 for some time that such ants bury their ma- 

 ture larvse in the soil till they spin their 

 cocoons and then, after some hours or days, 

 unearth them and cleanse the outer surface 

 of their envelopes. This habit has been repeat- 

 edly described for European species of Formica, 

 especially by Wasmann.* It is, of course, 



* See, e. g., his ' Vergleieliende Studien fiber 

 Ameisen- und Termitengiiste,' Haag, 1890, p. 95 



a necessary habit, as pointed out by Dr. Cook, 

 since the larva must have some support for the 

 threads spun from its sericteries. In default 

 of earth, as when ants (species of Campono- 

 tinse or Ponerinse) are kept in glass nests of 

 the Fielde pattern, the workers tear off bits of 

 sponge or toweling to spread over the larvse 

 as a support for the cocoon. If such materials 

 can not be obtained, or if the larvK are much 

 disturbed while spinning, they will form free 

 pupEB. I have seen this repeatedly 'in the case 

 of mixed colonies of our American amazons 

 (Polyergus hreviceps, P. hicolor and P. 

 lucidus) and their slaves, and in Odontoma- 

 chus clarus and Pachycondyla harpax (a 

 species allied to the kelep ant). The habit 

 has been emphasized by Wasmann because it 

 is of importance in his discovery of the inter- 

 esting relations of Formica to the parasitic 

 beetles Lomechusa and Atemeles: 



Both Lomechusa and Atemeles larvte are assisted 

 in their pupation by their hosts, exactly as are the 

 ant larvae. The adopted larva is placed on some 

 soft sand, and while it moves the anterior portion 

 of the body slowly to and fro, it is surrounded 

 with a ring of earth, which is soon built over it 

 like an arched roof. Within this the adopted 

 larva spins a very fine web which merely keeps the 

 surrounding sand particles together. 



This is the critical moment in the development 

 of the Atemeles and Lomechusa. The ants are in 

 the habit after a few days of unearthing their own 

 larvae which spin a dense, tough cocoon, in order 

 to cleanse and stack up the cocoons. They at- 

 tempt tlie same with the adopted larvae, unfortu- 

 nately, since they are thereby prevented from 

 pupating and perish, if they are not again buried 

 and left to themselves. Only such adopted larvae 

 develop, as are forgotten by the ants after they 

 have been embedded. For this rfeason I obtained 

 only one imago from thirty Lomechusa larvae and 

 not a single one from more than fifty Atemeles 

 larviE. 



In this connection it is perhaps timely to 

 call attention to the fact that C. Janet* some 

 years ago emphasized the bearing of the pres- 

 ence and absence of the cocoon in ants on the. 



et al., up to his recent article : ' Zur Kontroverse 

 fiber die psychischen Faliigkeiten der Ameisen,' 

 'Natur und Sohule, Bd. 3, 1904. 



* ' Les Fourmis,' Soc. Zool. de France, Conf. 28, 

 Fev. 1896, Paris, 1896, pp. 3, 4. 



