to have displaced a young cell or mesopore. Such a large acanthopore inflects the cell wall, forming a vertical rib or pseudoseptum as a rule. No doubt the acanthopore end, or wart on the wall, extended to fill an invagination in the web or cortex which bound the zooids to which the cells belonged. Explanation of the cause of such invagination need not be attempted here. But it may be added that the walls were evidently built by surface secretion, and that the growth of a projecting wart would be accumulative as compared to a plane surface, other things being equal. This may explain why acanthopores are

often so independently large. A scalloped or saw-edged wall would not develop so. The transverse of a wall would, however, if the zooids and cortex did not draw upwards fast enough to prevent it. Thus a wart thickens in two dimensions to the wall's one. Since, however, the wart is on the wall, it receives below the wall thickness also, and hence acanthopores might well be four or more times the thickness of the wall from result of secretion alone. In fact, it is necessary to explain why the walls, etc., can be very thin, which is evidently because the zooids drew forward rapidly.

It is fair to presume that structural differentiation of the wall may be accompanied by difference in composition; and that may explain why the elevated lunarial wall may be different colored and textured than the main wall; or, again, lucid vertical lines or "lunarial tubuli" only are different, these being the vertical extensions of tooth-like elevations on the lunarial wall margin. Acanthopores show similar differences. Such structures might be mistaken for mural pores in thin sections. Mural pores do not occur.

Tabulæ perhaps need no special mention; their character as the successive bottoms of the calycals is entirely evident (Plate A, Fig. 2, and Fig. 1, h, p. 21). In structure, if any be seen, they are one-sided. E. O. Ulrich has observed the very rare occurrence of perforation or circular opening in the last tabula, which he interprets to prove that all other tabulæ are double, comprising the amalgamated cover of one zoœcium and bottom of a next superimposed zoœcium. In absence of any substantiating evidence it is better to take the most direct explanation, which would be that the observed perforated tabulæ were left incomplete by the death of the colony. Indeed, in Hemiphragma, the tabulæ of the cells in the peripheral region are all left incomplete or were imperfect as to calcareous structure. The surfaces of tabulæ are sometimes papillated, and these again, like the acanthopores, simulate perforations in the fossilized specimen.

As a rule, neighboring cells do not have corresponding tabulæ, either in position or number. In any species or individual they are approximately regular in position and numbers, but never quite so. In different species the number ranges in extremes from none, as seen in the axial region of some species, to very many in others, or even to a compacted papillose mass filling the cells, or especially the mesopores in some thick-walled forms. There is no unit form and size of loculus assignable, which argues very strongly against any theory that the successive loculi represent superimposed zoœcia. The clearer interpretation is that each cell was built by one zooid. Tabulæ are, as a rule, very thin; and tabulæ wanting and tabulæ thickened are opposite degrees. Individual variation and specific difference in number and thickness of tabulæ may be ascribed to difference of growth and to secretion of substance. Difference in form, such as the cystiphrams (Prasopora), are ascribable to shape and size of the zooids' base; and vesiculose mesopores, to shallow or closed calycals from short zooids, and to their shifting, possibly, also.

AFFINITIES OF TREPOSTOMATA

Regarding the affinities of the Monticuliporoidea as a whole, the evidence uniting these to the Bryozoa on the one side, and to Tabulata or Alcyonarian corals on the other, does not lead to a compromise conclusion that they really were related to both as an intermediate or connecting link, because, as will be seen, the interpretation of the zoarium necessary to unite them with the one is discordant to that necessary to unite them with the other; and because of evidence to the contrary from the embryology of living Bryozoons and corals. Discussion is therefore confined to the question whether the extinct Monticuliporoidea are Bryozoa or Cœlenterata.

In relation to Bryozoa, the problem begins with the Trepostomata section, some or all of which have been variously and doubtfully referred to Cyclostomata; and this order of Bryozoa is extant. The reference involves comparison with the supposed Cyclostomatous genera Neuropora and Heteropora, of which we are yet uncertain. Gregory' refers these as typical Trepostomata, not Cyclostomata. The question rests mainly upon the fossil and recent Heteropora which Nicholson has thoroughly discussed and which, as it appears, simulates Trepostomata, but has many transverse mural pores and other differences. Trepostomata must therefore be proved to be Bryozoa and Heteropora likewise to belong to Trepostomata, before they can be united with assurance. Gregory's reference needs proof and affords no evidence, but expresses well perhaps that we are uncertain of all. Passing to the comparison of Trepostomata with undoubted Bryozoa, this requires knowledge of the extinct Cryptostomata, which must in turn be compared with Bryozoa; and discussion of that part of the problem will therefore be deferred to the section on Cryptostomata.

In relation to Tabulata or Alcyonarian corals, Trepostomata may be compared immediately. In the first place, such forms as Monotrypa compare with Chatetes, a massive zoarium of small, 'Catalogue of Jurassic Bryozoa, p. 193, 1896.

2 Structure and Affinities of the Genus Monticulipora, p. 62.

tabulated, thin-walled, polygonal cells. Chatetes has comparatively lighter colored, probably more calcareous walls. Its cells increase only by fission, while intermural "budding" obtains in Monotrypa. It is said, however, that fission occurs rarely in Monticuliporoidea, which leaves a difference in degree only between these two. Chatetes is extinct, but is referable only as a coral. It indicates that the Monotrypa, Monticulipora, etc., are corals; but that the family Chatetida should contain Monticulipora seems doubtful when Fistulipora is placed in a family of its own.1 Certainly Prasopora does not belong in both.

Fistulipora, which is the extreme form of Trepostome as compared to Monotrypa, is the very one most approaching the Recent coral, Heliopora, which, as shown by Mosely, is an Alcyonarian, with a true tabulate skeletal structure. Heliopora has the larger cells, but like Fistulipora has autocells among mesopores, called siphonopores. The autocells increase by "conenchymal gemmation," i. e., a young autocell arises among siphonopores, displacing several. Siphonopores and mesopores are alike. Also, in the mature region of Heliopora, the walls thicken and a wartlike projection stands generally at the siphonopore angle and twelve of them surround the autocell. Structurally the warts are similar to "acanthopores," but the wall of Heliopora is highly calcareous, and in thin section one sees primarily the crystalline structure radiating from the normal line or center, while the Monticuliporoid wall, being apparently less calcareous, shows the organic lamination, and the acanthopores have concentric structure. The difference is referable to the degree of calcareous deposit in which all corals differ. The warts on Heliopora are due to transverse canals between zooids swelling the cortex unevenly. Acanthopores might well be of like origin, and if so they indicate a canal system like that in Alcyonaria.

It has not been clearly enough understood that Mosely 2 demonstrated the Heliopora to have no mesenterial septa, but that twelve vertical ribs in each autocell are pseudosepta; and EASTMAN, op. cit., pp. 102, 103. 2 Challenger Report.

as such they can be compared exactly with the inflections produced by acanthopores in many Monticuliporoids, if the small difference in calcareousness of skeleton be considered. More calcareous ones have sharper processes. A lunarium is wanting in Heliporida, but this structure is absent in Fistulipora in part, and in most Trepostomata. Any structural differences between Heliopora and Fistulipora are found further in some genus or other closely related to the former, except the monticules or maculæ of the latter.

If one places all Trepostomata and the Tabulata (Alcyonaria) together, they are compared as follows: The largest cells of the former are scarcely equal the smallest autocells of the latter. Growth habit is alike. Cells, monomorphic or dimorphic, are alike; except that distinct pseudosepta in autocells are common in Tabulata, being absent in few cases and these when the walls are very similar in structure to that of Trepostomata, i. e., when crystalline radiate striping is absent,except also that mural pores cross the walls of many Tabulata, not, however, in dimorphic forms nor in monomorphic ones with small cells, except again, the lunarium of Fistuliporidæ and the so-called dorsal septum of Alveolitidæ. Notably, the two lunarial angles, forming two pseudosepta, are on the upper side in the former, the single "septum," pseudoseptum, is on the lower side in the latter; the structures therefore not corresponding. If, however, they be ascribed respectively to the double ventral folds and the single dorsal fold of certain Alcyonarian Recent corals, one finds them all represented in Conites, a Paleozoic tabulate coral. They are rarely indicated in skeletal structure of either Monticuliporoidea or Tabulata, but argue Alcyoranian affinities.

Cell increase is not unlike. In the Tabulata (Alcyonaria) it is by fission, unequal fission, stolonal gemmation, and intermural gemmation, which are probably degrees of transformation.2 Intermural gemmation is the rule in Trepostomata. The peculiar 'See further, Neues Jahrb. Minn. Geol. and Pal. Beilb. X, pp. 316, 320.

2 See op. cit., pp. 281, 359.