FAMILY 10. PURPURIFERA. 231 
Linneus and the earlier writers had so few opportunities of obtaining 
* the soft and living parts of the Mollusca, that they were contented to 
classify them according to the external form and development of their 
shells. The distribution which they adopted in this emergency, being 
nevertheless executed with judgement, has proved in some instances to 
have been tolerably correct ; but their method of associating in one divi- 
sion all shells of a tubular construction did certainly involve a most 
singular anomaly of organization. It may have been already noticed how 
perfectly distinct from each other are the Serpule, the Aspergilla, the 
Vermeti, the Dentalia, and the Siliquarie ; and we shall now show that the 
Magilus differs as essentially as any of them from the character that was 
at first attributed to it. The shell of Magilus has indeed most singularly 
taxed the ingenuity of naturalists to discover the nature and affinities of 
its animal inhabitant; for whilst De Montford, when establishing the 
present genus for its reception, included it with the Mollusca, Lamarck 
and others referred it to a place amongst the Serpulaceous Annelides. It has 
been classified with the Mollusca, however, only on account of an imaginary 
affinity with the Vermeti; no one could have suspected it to be an isolated 
and extraordinary modification of a true pectinibranchiate gastropod, as 
lately discovered by Riippell. The Magilus was found by that indefatigable 
Abyssinian traveller on the shores of the Red Sea imbedded in a particu- 
lar kind of madrepore, and its peculiarities minister in a surprising man- 
ner to the nature of its existence. The formation of the shell commences 
in the same style of volution as the rest of the Purpurifera ; but, that the 
animal may retain free communication with the surrounding fluid, and 
keep pace at the same time with the increase of the madrepore, it raises 
itself by depositing a sufficiently abundant secretion of calcareous matter 
to completely solidify the early portion of the shell. The mollusk and its 
shell then leaves the spiral plan of construction, advancing in a straight 
or flexuous direction with the increase of the madrepore, and, as the re- 
volved portion of the shell is immoveably imbedded, the new method 
of growth could only be effected by this extraordinary effusion of calca- 
reous matter. The soft parts of the Magilus do not exceed an inch and 
a half to two inches in length, but the shell varies from two to fifteen 
