laterally, but left the anterior tergal plates intact. The result is shown photographically in Figure 205.

Specimen In. 31268-Figures 35, 53, 54, 76, 206, 207, 209 and 214. (This specimen is marked also No. 53 of the W. Egginton Collection in the British Museum.) This specimen is of the greatest importance for the understanding of the arrangement of the coxae and of the segmentation of the abdomen. It is unusually well preserved and three legs are complete, so that of the four pairs only the last cannot be measured, both legs being preserved only as far as the end of the femur. Piece a showing the dorsal surface is reproduced photographically in Figure 211, Piece b showing the ventral surface in Figure 212. The segmentation of the abdomen is shown in Figures 53 and 54, and the two surfaces superimposed in two colors in Figure 35. Figure 76 shows the ventral aspect of the specimen with missing appendages restored and represented in dotted lines. For some reason unknown to me, this specimen does not show the "arched area" and the latter is therefore omitted from the drawing. The genital region is shown photographically in Figure 214. It is extremely well preserved, although slightly pressed out of shape. Depression A leaves no doubt that it is a female. At the sides of the depressions the three pairs of lungs are completely and beautifully preserved together with their spiracles and the apodemes for the pulmonary muscles. No less beautifully preserved are the coxae, including all six pairs. The cheliceral coxae (Figure 209) happen to be rather widely open, showing the four foramina to their advantage. The outer walls of the cheliceral coxae partly obstruct the view of the pedipalpal coxae, which appear therefore as narrow slits. The wide mouth may be seen behind these four coxae, and behind the mouth a smooth trough leading to the end of the cephalothorax. The trough is either the foregut itself, or else the endosternite underlying it. Total length 10.14 mm. Carapace 4.29 mm. long, 4.00 mm. wide. Abdomen without the first tergite which is not visible, 5.86 mm. long, 5.71 mm. wide, measured to middle and the measurement doubled. Median tergal area 2.86 mm. long, 2.86 mm. wide. Ninth tergite 1.00 mm. long, 1.43 mm. wide. Angle formed by the sides of the eighth sternite ca. 120°.

The legs are amazingly alike and only 1.2 times as long as the carapace, except the fourth leg, which must have been longer if judged by its femur.

Specimen In. 15858-Figures 73, 200 and 202. This specimen is also very well preserved. Its first leg is complete, the others preserved to the end of the patella. Consequently it was possible to make a good cast of plastic clay and to photograph it in threequarter view (Figure 202) to show the shape of the body and the general appearance of the animal. The crest is shown also photographically in Figure 200 and a drawing of the genital area is given in Figure 73. The specimen is a female of approximately 12 mm. length. Carapace 4.57 mm. long, 4.00 mm. wide. Abdomen, including the first tergite which is plainly visible, 7.14 mm. long, 6.00 mm. wide. Median tergal area 3.57 mm. long, 3.14 mm. wide. Ninth tergite 1.00 mm. long, 1.57 mm. wide. Angle formed by the sides of the eighth sternite ca. 120°. First leg, femur 1.86, patella 1.00, tibia 1.00, metatarsus 0.57, tarsus 0.86. Total 5.29 mm. Second and third femur 1.86 mm. Fourth femur 2.14 mm.

Genus Coryphomartus Petrunkevitch, 1945

Type C. triangularis (Petrunkevitch), sub Anthracomartus The type of this genus, represented by a single specimen from the Upper Coal Measures of Joggins Mines, Nova Scotia, was described by me in 1913. The confusion of lines made me give a wrong interpretation of the dorsal surface. This time I was able to study the specimen once more, to clean it and to give new, correct figures. The genus is represented by a single species. The characters of the genus are given in the key. The specimen is shown photographically, as it appears now, in Figure 191. Figure 59 shows the carapace and the abdomen with the lines corrected and the segments numbered in conformity with the other specimens of Anthracomarti. The median tergal area formed by tergites 4 to 8 is 2.66 mm. long, 2.57 mm. wide.

Genus Brachypyge Woodward, 1878
Type B. carbonis Woodward

The characters of the genus are given in the key. The segmentation of the abdomen is shown in Figures 60 and 61, based on the original

drawing in Woodward's paper, and conforms with the other species of Anthracomarti. For the original description of the specimen, the reader is referred to Woodward, H., Geol. Magazine, (2), Vol. V, 1878, pp. 433-436, Plate XI. For a photograph of the type see P. Pruvost, Mem. Mus. R. Hist. Nat. de Belgique, No. 44, 1930. Plate XI, figures 1 and 2. Additional data concerning the species on pp. 210 - 212.

Genus Maiocercus Pocock, 1911

Type M. carbonis (Howard and Thomas), sub Eophrynus.

The characters of this genus are given in the key. As long as the genus is recognized, the type species is, of course, "carbonis" and not celticus as given in my Monograph of 1913. Pocock had to find a new name, because in 1902 he referred this species to Woodwards genus Brachypyge, the type of which is a species which happens also to have the name carbonis. Pocock retained the name celticus in 1911 possibly out of oversight. Pruvost recognized the genus Maiocercus in 1911 when he referred to it what he thought to be a new species, Maiocercus orbicularis Gill. Of this species Pruvost gave excellent photographs of two French specimens on Plate XXIII and a detailed description with two line-drawings (pages 359-363). In 1930 Pruvost dropped the Genus Maiocercus, synonymizing it with Brachypyge and the species M. carbonis Pruvost (nec Gill) with B. celtica Pocock (p. 213). He gave a photograph of a Belgian specimen on Plate XI. I have explained above the reasons why I retain the Genus Maiocercus here. In Figures 62 and 63 I have reproduced Pruvost's figures of 1919, but with the introduction of segmentation conforming with the one adopted in the present paper.

Evolutionay trends in Anthracomarti

The ideas concerning evolutionary trends in Anthracomarti, which I put forward in 1945 have to be considerably modified by the discovery that the Order as universally recognized at that time, contained definitely unrelated genera. Now that the Order Anthracomarti is restricted here to the seven closely related genera, the obvious fact emerges that all seven have the same fixed number of abdominal segments. They also have the same number of book-lungs and the same number and type of structures in the genital region, at least in so far as four species of two genera may be adduced in evidence. That leaves only modifications of the carapace, which I have used for the

separation of genera. Of course, the chief character of the new Subclass Stethostomata which I proposed for the inclusion of Anthracomarti and Haptopoda, is in itself evidence of an important trend which reached its climax in these two Orders. Perhaps the fact of their extinction was due to the inability of the abdomen to undergo any change involving segmentation, and to the impossibility of returning to a less specialized arrangement of the coxal region.

Introductory Remarks.

III. SUBCLASS SOLUTA

This Subclass, as its name implies, is proposed here for the inclusion of a single Order Trigonotarbi, in which certain characters were still in a labile condition when its representatives faced extinction. The chief of these characters was the manner of juncture between the cephalothorax and the abdomen and the consequent fate of the first abdominal segment. We have seen that in Latigastra there is a trend to lose the first segment. Sometimes the entire segment disappears, at other times the sternite is normal, but the tergite is either shortened or fused with the carapace. In Caulogastra the reduction of the first segment is accomplished through peripheral constriction, and not by total shortening. It reaches its climax in Araneae where it is reduced to a very slender pedicel. Now in Trigonotarbi the behavior of the first segment and the manner of juncture between the abdomen and the cephalothorax differs in different genera. In Trigonotarbus the juncture occupies the full width of the abdomen and the first tergite is somewhat abbreviated. In Anthracosiro the juncture is also full width, but there is a distinct, though not considerable constriction between the carapace and the abdomen, and the first tergite is represented only by the soft connecting membrane with its dorsal muscle fibers. In Pleophrynus the first tergite is as wide as the carapace, but it is connected with the latter only by its middle third. Eophrynus has probably the same type of connection, while in Palaeocharinus, judging by Hirst's Figure 4, the connection most resembles that described by me in the present paper for species of Anthracomarti.

The second important character is furnished by the segmentation of the abdomen. Here a distinct trend may be observed toward a loss of posterior segments. Palaeocharinus, according to Hirst, has an abdomen composed of eleven segments the last two of which are reduced to little, short cylinders or rings forming a pygidium.

Anthracosiro, as we shall see, has ten abdominal segments, Eophrynus and allied genera have nine segments, Trigonotarbus has eight segments. Trigonomartus, with related genera, has also only eight segments, but there is a difference in the structure of the last segment in Trigonomartidae as compared with that in all other Trigonotarbi. In Trigonomartidae the last tergite has a pair of marginal fields or plates, in the other genera it has none. Instead, the single plate of which it is composed is bordered in front by the median plate of the penultimate tergite and on the sides by the marginal plates of the penultimate segment, an arrangement made possible by the great curvature of the tergite. In these genera the last tergite is also bent under in the same sense as in the Anthracomarti, appearing on the ventral side of the abdomen. Here the sternites have no marginal plates, but the penultimate sternite is bent almost as a hairpin, enclosing in the space between its two arms the last sternite with the anal operculum and the ventral portion of the last tergite. A distinct line separates the tergite from the sternite in Eophrynus and Pleophrynus. In Anthracosiro the line is faint, but visible. I have indicated it in Figure 114 by a broken line. In Trigonotarbus (Figure 111) it is very faint in three specimens and not visible in the others, this being the reason why Pocock overlooked it. On the other hand in Trigonomartidae the last tergite does not extend unto the ventral surface. The last sternite is biconvex and occupies the entire end of the ventral surface of the abdomen.

8. Order TRIGONOTARBI, new
Characters of the Order

Fossil Arachnida in which the abdomen is joined to the cephalothorax either by the full width of the carapace or only by a median portion of it, while the width of the first tergite remains the same as that of the carapace. Carapace entire, of various shapes, often with a pair of eyes. Abdomen composed of from eleven to eight segments, not counting the anal operculum. The latter is situated on the last sternite and has an anterior and a posterior valve. The tergites are divided into three longitudinal rows, so that each tergite has a single, median, and a pair of lateral, marginal, plates. The last tergite either is subdivided like the other tergites or has no marginal plates, is surrounded by the penultimate tergite and extends to the ventral surface, where it is wedged in, together with the last sternite, between the arms of the penultimate sternite. All