6 BULLETIN 93, UNITED STATES NATIONAL MUSEUM. 



In most recent barnacles the number of compartments lias been 

 reduced, and it becojnes important to determine the homologies of 

 those remaining. This is partly to be done by attention to the 

 articulating edges of the compartments. 



The carina and rostrum are nearly alike in primitive forms, both 

 being bilaterally symmetrical, and having alse on both sides, therefore 

 overlapped by the adjacent compartments. There is no certain 

 means of distinguishing them apart when detached, but the carina 

 is usually more narrowly curved or bent than the rostmm of the same 

 species. 



The carinolateral and lateral compartments are asymmetrical, 

 having alx on their rostral edges, and radii, or an overlapping border, 

 on the carinal edges. There is no significant difference between these 

 two compartments when detached, though the lateral is wider in most 

 genera. 



The rostrolateral compartments have no alse, both sides over- 

 lapping the adjacent compartments, thus differing essentially from 

 all of the other plates of the wall. 



Departures from the structure just described are due to concres- 

 ence of compartments. Reasoning from the conditions in species 

 still in transition stages, we niray infer that reduction in the number 

 of compartments has taken place by two modes: 



1. Reduction by elimination. In Acasta and Conopea the carino- 

 lateral compartment is usually very nan'ow. In Acasta sporillus it 

 has been almost crowded out, and does not reach to the basis. In 

 Conopea cornuta it has entirely disappeared. In some other genera 

 this compartment appears^ to be wanting.^ 



2. Reduction by concrescence. In Chelonihia the rostrolateral 

 compartments and rostrum are calcified together, but traces of the 

 sutures are usually discernable. In Pachylasma crinoidophilum the 

 same compartments are united by fine Ihiear sutures, but can be 

 separated; in P. darwinianum they are calcified together externally, 

 but the sutures are visible inside; finally in P. giganteum these com- 

 partments are united by sutures in the earliest sessile stage, but 

 aftenvards they become completely concrescent, so that no trace 

 whatever of the tripartite nature of the compartment is visible. The 

 result of this concrescence is that the composite rostrum (rostrum plus 

 rostral latera) has radii or overlapping borders on both sides, these 

 being the original overlapping borders of the rostral latera. Havin'g 

 this transition series, among others before us, it seems highly probable 



' In such genera as Tctradila and Chlhamalu^, in which tlie carinolateral compartments are absent, 

 they may be fused with the lateral comE,artments or with the carina; but seeing tliat they are normally 

 developed later than the other valves, It appears to be the simplest theory to assume, milil the contrary 

 be proved, that they are aborted. Finally, the somewhat unexpected conclusion that the shell (not includ- 

 ing the operculum) of sessile cirripedes normally consists of eiglit valves^four belonging to an upper whorl 

 and four to a lower whorl, all forced into a siTiglc ring, and often more or less fused together, though not 

 strictly proved, is rendered highly probable. (Darwin.) 



