BULLETIN 59, UNITED STATES NATIONAL MUSEUM. 175 



Botweon a paliis and the wall there normally is one traheciila. whicli tcniiinates 

 on the .septal margin in a septal granule, slightly removecl from llic wall and of 

 moderate prominence. Near the lower edge of the living portion the ealices are 

 shallower, and rather often there are two granules on the septal margins between the 

 pains and the wall. The trabeculte next the wall are very rarely eonneeted by a 

 complete ring of synapticula, but one is nearly always paitialiy rejjresenti'd. The 

 synaptieula may be free from the wall or more or less fused with it. There is no 

 complete and persistently developed mural shelf between the septal granules and the 

 wall, but one is often present in parts of ealices. In thickness the sei)ta e(pial or 

 somewhat exceed the width of the ititerceptal loculi; the outer ends are thicker than 

 the inner portions. 



The pali are moderately wide and rather tall; the fornuda is usually complete, 

 but the palus is often small and sometimes absent on the ventral directive. They are 

 joined together bj' a palar ring of synapticula, which is not always complete. 



The columella consists of a compressed, thin, lather prominent tut)ercle, con- 

 nected by 6 more or less definite I'ays to the inner ends to the septa. The col- 

 umellar tangle is often considerably compacted by stereoplasmic deposit. 



The skeletal surfaces are covered b}' low. blunt, somewhat crowded granidations. 



The preceding description is based on Dana's type, as has already been stated. 

 Forty -six specimens and a few fragments belonging to the United States National 

 Museum are referred to /*. comprenKd. Of these. 3 have belonged in the institution for 

 sometime; 7 were collected l)y the ^l/i'/^rw*- expedition of l'.ti)'2; 'MS were received 

 from Dr. W. T. Brigham, of Honolulu. These specimens present a wide lange of 

 vai'iation, so wide that it was very difficult to devise a scheme l>v which tiie facts 

 obtained from studying th(>m could be intcliigil>ly placed on record. Sixteen ])rincipal 

 types and 4 subordinate tvpes of variation are recognized. The vaiialion appears to 

 be continuous, but with a number of definite secondary modes, should they l>e plotted 

 into a specific curve. What the piiysiologic meaning of this \ariation is. it is at 

 present entirely impossible to say. ^\'c ha\e no facts bj' which itcould beascertained 

 whether the difl'erences are of gametic oi vegetative origin. It was therefore 

 decided to designate the different types of variation represented as '" forma'." the sub- 

 ordinate types as "subformffi" of "forma'." Latin names are attached to the forms 

 and subforms. They are intended principally as descriptive terms to enable writing 

 or speaking of the types of variation, and pn)bal)ly will not be considered of jxirticular 

 importance except bj* those who are making detjiiled studies of varialion. 1 ])ri'fer 

 naming to Bernard's number system. 



Before proceeding to a consideration of the variations of the species, a statement 

 of the fundamental plan underlying all of them will be made. 



SjHrljie dt'serij)tion of P. coi/ijircKm.- The corallum has a slightly expanded or 

 incrusting base, above which it soon becomes ramose, producing crests or branches, 

 usually more or less compressed. The branches near the base may be much fused 

 one to another or free; the same remark applies to the condition in the upper 

 [lart of the corallum, the branches, plates, or crests maybe separate or fused into 

 plates of varying width. The terminals or upper edges arc aiii.Dst iinaiialiiy 

 truncate; the ends of the branches are frecjuently clavate. 



