UMBELLULIDAE 275 



the same sexual condition, living in the same locality in deep water, should belong to more than one 

 species. Any cross-fertilization that might occur would soon swamp any specific distinctions.' This of 

 course to a great extent depends on what we consider as specific characters. Nearly related species 

 may according to the laws of heredity live and propagate side by side and nevertheless keep their 

 specific identities intact, and so far as we know this is as true for the deep sea as for every other domain. 

 The Indian Ocean contains an extraordinarily rich fauna in its deeper regions, but at present we do 

 not know, which factors, or combination of factors, make this ocean especially favourable for 

 Umbelluh and for the moment it must suffice to record the known facts. 



U. spicata appears to be confined to the Indian Ocean and at the same time has a remarkably narrow 

 bathymetrical range between 344 and 1281 m. Different growth stages have been interpreted as 

 distinct species, and because of the great flexibility of the stem, adult specimens had not been caught 

 and described hitherto. 



U. pellucida also seems to be restricted to the Indian Ocean. Its known bathymetric range is from 

 220 to 2001 m. It is noteworthy that the deepest haul only gave one specimen, whereas the haul in 

 220 m. resulted in some 240 specimens; also one haul in 625 m. depth gave 8 specimens, all other 

 catches consist of i or 2 specimens. This may be taken as an indication that the upper part of the 

 habitat oflters the optimum conditions, and that the deepest locality, separated widely from the 

 next record, shows an erratic individual which has grown up below the normal habitat of the 

 species. 



U. lindahli is the most widespread species of the genus, and if U. encriuus turns out to be an eco- 

 logically distinct form, the species inhabits all oceans. At the same time the bathymetric range — 

 77~S3i° m. — is much greater than in the other species. Most of the earlier records in the Antarctic 

 region were from great depths, from 2450 to 5310 m., and only two localities showed depths as shallow 

 as 540 and 457 m. It was therefore surprising that all the nine Discovery Stations lie between 645 

 and 77 m. 



It has been rather usual to regard temperature as the deciding factor, when deeper living bottom 

 animals ascend in higher latitudes to lesser depths than elsewhere. U. lindahli, on the other hand, 

 is remarkable in that both the deepest locality (5310 m.), and the shallowest (77 m.) are situated in 

 Antarctica, and that the deepest record here almost coincides with depths in the Atlantic near the 

 equator. 



One may say that generally the more robust and better developed specimens have been observed in 

 Antarctica, the slenderer specimens in deeper waters of the tropical seas. However, there are excep- 

 tions, e.g. the comparatively short and fleshy 'antarctica ' specimens are common both in the Antarctic 

 and in the tropical part of the Indian Ocean. The question of geographically distinct varieties in 

 U. lindahli must also be considered. 



In spite of comparatively numerous localities (cf. text-fig. 4, 49 records, 16 of which are in Antarctic 

 waters) the evidence is scanty, and it is not astonishing that U. lindahli has been described under several 

 names, which in some cases might seem to indicate local races. The Discovery collections have 

 proved that the three ' species ' only described from Antarctic waters represent either developmental 

 phases, or more or less luxuriant individuals. ' U. carpenteri ' doubtless represents younger stages in 

 most cases and larger specimens cannot be distinguished from ' U. magniflora (ambigua) '. Nothing 

 indicates that these forms are restricted to special geographical areas. A more luxuriant development 

 of the cluster leads from U. magniflora to U. antarctica ; from the specific name one might assume that 

 the latter represented a special Antarctic variety, but typical U. antarctica are also comparatively 

 frequent in tropical Indian waters. 



The accessible data do not yet indicate geographical races, or varieties, in U. lindahli. On the 



