^ tevisSon of H.Tyrmn'oda 5 Arachnida ; ^ ^ p^p'^YM Cl i'^ Q.'rioioqy it' *;.;Tr/ S.‘K‘ai..o;jo.i Mu! /"i’. Bi'W eum A SERIES OF OCCASIONAL PAPERS ON DIPLOPODA AND OTH ER: ARTHROPOD A -r-rf- - By'.T;T;4i:feT^' :vA- ■' O. F. COOK. Huntington, New York, 1896. - T, a'. T- /' ? ]'i>l' / i i\, i T '^':T I') I /'I I. A SYNOPSIS OF MALAYAN PLATYRRH ACID^, A recent opportunity of examining for the second time some of the types of species of this group in the Berlin Museum has confirmed the opinion that several genera have hitherto been confused under the names Stenonia, Platyrrhacus, and Acanthodesmus. As a begin- ning toward the elucidation of the group the following synopsis of generic types studied by me is here offered. Drawings and extended descriptions are in preparation. Very lar^e (90-130 mm. in length) ; dorsum very slightly granular, coriace- ous, or smooth, except that there are three more or less distinct transverse rows of small tubercles ; lateral margin of carinse with several coarse teeth ; first segment broadest in front, the anterior corners produced and rounded : Genus Phyodesmus, type Ph. pictus (Peters), Borneo. Small to large (not exceeding go mm.); dorsum densely granular, or the lateral carinae with margins subentire, erose, or deeply bifid ; first segment subelliptic, broadest at or behind the middle, the anterior corners not de- veloped. Sterna with two pairs of long, sharp spines, of which the anterior pair is di- rected cephalad, the posterior caudad ; carinse laterally margined outside the pore : Genus Phractodesmus, type Phr. subvittatus (Peters), Linga. Sterna unarmed, or with four more or less distinct, short, conic spines all di- rected ventrad or slightly caudad ; carinse laterally immarginate, or the pore included in the margin. ........... Copulatory legs unbranched, extended into a long, coiled process : Genus Derodesmus, type D . flai^elUfer, sp, ??., Ternate. Copulatory legs with two or three distinct prongs. ..... Dorsum with three conspicuous and equal transverse rows of prominent, oval, smooth, tubercles ; carinse laterally strongly margined ; margin smooth and prominent, below with a large rounded tooth ; pores borne in lateral exca- vations of the thickened margins, except on posterior segments where they are subdorsal : Genus Taphodesmus, type T. moluccensis (Peters). Dorsum with the posterior row of tubercles much more prominent than the usuallv inconspicuous anterior rows; lateral margins of carinse not thickened nor smooth. ..... ......... Copulatory legs with the horn3^ transparent and smooth ungual portion of the last joint about twice as long as the oval basal hairy part; basal parts of the two legs in situ approximate ; distal portions curved mesad at base and crossing each other, after v/hich a broad, short, flat incurved lamina is given off the dorsal (anterior) side, v^hile the long slender ramus is again bent mesad and recurved upon its fellow, so that a large oval space appears to be entirely enclosed between the terminal rami of the two legs when viewed in situ from below: Genus Ilodes7nus , type /. (Brandt), Manila. Copulatory legs with the ungual rami not longer than the hairy basal por- tion, and not crossing each other before branching; also not recurved Brandtia, p. I. 2 PLA TYRRHACIDA£. upon each other so as to enclose a median space ; the shorter anterior (dorsal) ramus is moreover not conspicuously flattened or expanded. .... Copulatory legs with basal hairy portion of the last joint very narrow and slender, nowhere broader than the ungual portion at the base of the first (lateral) division, which is long, slender, at first nearly straight, and almost perpendicular to the basal part of the joint; the mesial ramus is at first flat and rather broad, crosses its fellow, and near the apex is again divided, the three branches being short, subequal and divaricate; Genus Psaphodesmus, type Ps. concolor (Peters), Moti. Copulatory legs with hairy basal portion of last joint conspiculously thicker and broader than distal horny part. ........ Body rather slender, and with the carinse rather remote ; dorsum scarcely convex, nearly smooth, or indistinctly granular laterad ; carinse with lateral margins entire or faintly sinuate ; corners rounded, the posterior produced only on posterior segments ; pores removed from the margin by more than the width of the poriferous ring: Genus Leurodesmtis, type L sumatranus (Peters). Body robust and with the carinse adjacent, or the dorsum distinctly convex and distinctly granular ; carinse with lateral margins dentate, incised, or the posterior corner distinctly angled or produced on middle segments ; pores variously located. ............ Last segment subtriangular, the apex distinctly narrowed and produced ; two prominent setiferous tubercles near the longitudinal middle of the dorsal surface of the segment; carinse with anterior margin straight, evenly and distinctly dentate ; anterior corner of canna prominent, sharply angled or with a distinct and prominent tooth ; lateral margin distinctly and evenly den- tate ; posterior corner produced on middle and posterior segments into a spur- like process which projects caudad but not laterad beyond the general line of the lateral margin ; dorsum strongly convex ; first segment ecarinate ; copula- tory legs not approximate, conic, rather short, branched immediately above the hairy basal portion ; terminal rami subequal in length, the posterior nearly straight, the anterior turned mesad to meet its fellow and decurved so that the apex is nearly in contact with the base; Genus Mniodesmus, typed/??. crossotus, sp. n.y Java. Last segment broad and subquadrate, the apex not narrowed ; truncate or very broadly rounded ; setiferous tubercles not prominent, and not remote from the posterior margin ; carinse v/ith anterior margin somewhat rounded and en- tire, or straight, with very numerous fine irregular teeth ; anterior corner rounded and without a prominent tooth or angle ; lateral margin sparsely and irregularly dentate, sinuate or subentire ; posterior corner slightly produced or with a large spur-like process projecting caudad and laterad beyond the ante- rior part of the lateral margin ; dorsum slightly or moderately convex ; first segment more or less carinate, that is, with a more or less developed lateral pro- cess ; copulatory legs more or less approximate, the anterior ramus not re- curved Transverse constriction (suture) costulate (crenulate) ; pores adjacent to the margin ; lateral margins of carinse sparsely and irregularly dentate and appa- rently deeply excised, from the fact that the long spur-like process of the pos- terior corner extends laterad as well as caudad ; carinm rather narrow, about one-third as wide as the body-cylinder; dorsum with three distinct though in- conspicuous rows of tubercles: Genus Xerodesmus, type X, drains, sp. n., Java. PLA TYRRHACID^. 3 Transverse constriction not costulate ; poriferous ring distant from the margin by a space equal to its diameter ; lateral margins of carinae slightly and broadly sinuate, not excised; the posterior corner square, not produced into a spur-like process ; carinse broad, equal to half the body-cylinder ; dorsum densely and evenly granular, the tubercles obsolete : Genus Acanthodesmus Peters, type A. pilipes (Peters), Borneo. The types of Acanthodesmus pilipes are at Berlin, but both are females. The character of the copulatory legs has accordingly been inferred from A. andersonii Poc."^, which appears to be at least con- generic with pilipes. The preceding table by no means exhausts the Oriental genera of this group as represented more conspicuously in the various papers of Mr. Pocock. Some of these can, it would seem, be safely indicated from that author’s excellent descriptions and plates, and the following table may be looked upon as supplementary, being intended to point out the differential characters of some of the forms which present the more striking divergencies from the genera established above. Copulatory legs with three prongs ; sterna not spined ; lateral margin of Carina with 4-6 rounded tubercles ; posterior corner of carina not produced into a sharp spur ; anterior and posterior margins of carinse smooth : Genus Zodesmus, type Z. tuberosus (Poc.t), Ki Islands. Copulatory legs with two prongs, or the sterna spined. .... Lateral carinse deeply bidentate, the pore located at the base of the anterior tooth: Genus Dicrodesmus SilvestriJ, type D. bide?is (Poc.), Sumatra. Anterior segments, at least, not bilobed, the others not deeply so, with the lobes and sinus more or less denticulate. ....... Carinse projecting obliquely forward so that even on the seventeenth seg- ment the posterior corner does not project farther back than the posterior margin of the segment : Genus Prodesmus, type Pr. stcbmissus (Poc.), Sumatra. Segments in front of the seventeenth with the carinae curved or produced farther caudad than the middle of the segments. Sterna spined ; copulatory legs with three prongs not greatly unequal in length : Genus Cradodesmus, type Cr. subspinosus (Poc.) Sterna not spined ; copulatory legs with two prongs. ..... Carinse of middle segments deeply and broadly excised ; the pore located close to the edge ; first segment with carinae short, directed backward and up. ward; dorsal surface nearly smooth: Leucodesmus, type L. weberi (Poc.), Sumatra. Carinse not excised, the poriferous ring remote from the edge by about its own diameter ; first segment with carinse longer, not directed candad ; dorsal surface densely granular: Genus Harpodesmus type H. laticollis (Poc.), Sumatra. *Ann. Mus. Civ. Genova (2) XIV, p. 788, 1895. t Ann. & Mag. Nat. Hist. (6) XV, p. 131, PI. IX, figs. 3-3 b, 1893. The other species of this table are to be found in Weber’s Reise, pp. 345-357. f Ann. Mus. Civ. Genova (2) XVI, p. 190, 1896. This paper came to hand after I had sent the same genus to the printer under another name. Two other geneia are proposed in the same place: Cyrtorrhachis, type Platyrrhacus sub- albus Poc., and Acisternum, type Platyrrhacus mo7i.ticola Poc., the first from Java, the second from Sumatra. 4 PLA T YRRHACIDAl. Under Pliyodesmus will be included, for the present at least, all the large forms related to P. pictus, such as P. inirayidus (Poc.), P. pfeiff- ercB (H. & S.), P. princeps (Gerv.), P. 7}iagnificus (Silvestri), and two new species in the Berlin Museum, P. petersii (No. 268) and P. mo?itrado (No. 267). These last were among the specimens indi- cated by Peters as types of pictus, from which both differ in their smaller size and shorter copulatory legs with a broader ensiform prox- imal ramus and a large process from the longer arm. From each other they differ in that petei^sii is more slender and with the branches of the copulatory legs much longer in proportion to the proximal hairy portion of the last joint The genus Phractodesmus is distinguished in the possession of ex- ceedingly large sternal spines, while its other characters are not re- markable. The “types” of Psapliodesmus concolor also numerous and in- clude several species. Among them was found the specimen (No. 280) upon which was established the genus Derodesmus. The cop- ulatory legs of this last are peculiar in the long slender last joint, hairy about half its length, with the terminal portion entirely un- branched, but coiled nearly in a circle. Dorsally the specimen is sculptured much as in coiicolor, but it is much smaller and less convex. As the true type of co7icolor may be taken the first specimens indica- ted by number (241). The specimens from Dodiiiga (No. 243) are much larger in size and nearly black in color, even to the copulatory legs. These it is proposed to describe as PsapJiodes7nus dodinga. The genus nearest related to Psaphodesmus is probably Zodesmus, which supposition is in accordance with what little is known of the geographic distribution. The type of Ueurodesmus has, as Peters de- clared, much external resemblance to certain South American forms placed under Rhacophorus, but it is without doubt a member of the present family. It is perhaps too early to attempt the subdivision of the Platyrrha- cidae into groups of higher than generic rank, and yet it is evident that there is little close relationship between the Malayan and South American forms. Especially is this the case in such instances as Phyodesmus, Taphodesmus and Psaphodesmus, which can safely be recognized as types of at least subfamily importance. The Phyodes- mini and Taphodesmini include as yet but a single genus each^ Taphodesmus, though known only from a fragment, is so peculiar that further specimens may easily justify the establishment of a distinct family. To the Psaphodesmini may be referred Derodesmus and Zo- desmus, while the remaining Malayan genera can be provisionally included in the Acanthodesmini, from which the separation of other subfamilies will doubtless be found convenient. May 7, 1896. 11. ON RECENT DIPLOPOD NAMES. An attempt at an arrangement of the diplopod families and genera was published in the preceding year as an introductory note to a revision of the American Craspedosmatidae. Apparently without having seen this paper Sig. Filippo Silvestri has recently prepared t a similar list in which many new families and genera are proposed, several of which were already provided. Other genera not included in ray arrangement have been published by Pocock, Verhoeff, Broelle- mann, and Porat, all of which appear in Sig. Silvestri ’s alphabetical list with the exception of four of Mr. Pocock ’s genera of Zephronii- dae t. Of this list the following names are believed to be una- vailable, or exception is taken to their disposition by some of the writers mentioned. Aporodesmus Porat, 1894. The type of Pocock’s genus of the same name is A. vicentii Poc., from the West Indies, and this is in all probability not congeneric with A. gabonicus (Lucas), whatever that may be, which stands as the type of Aporodesmus Porat. Moreover, Porat ’s genus appears to be older than Pocock’s by several months, so that the name must be reserved for the African form ; accordingly vmcentii of Pocock may be given the new generic name Docodesmus. Attemsia Cook, 1895. Since recognizing Verhoeff’s subgenus as worthy of generic rank, I have seen specimens from the Adelsberger Grotte, in all probability referable to A. stygia (Latzel). They were not congeneric with the types of Craspedosina rawlinsii in the British Museum. Boeemania Silvestri, 1896. Unfortunately, this name has been used for a genus of fishes ; for the genus of Lysiopetalidae may be substituted Trypostrephon, nom, nov., with T. ojientalis (Silvestri) as type. ♦Ann. N. Y. Acad. Sci. IX, pp. 1-4, October 1895. t Ann. Mus. Civ. Genova (2) XVI, March 1896. t Zoosphaeriurn, Arthrosphaera. Castanotherium, and Cyliosoma, Ann. & Mag. Nat. Hist., (6) XVI, Nov. 1895, pp. 410-414. Brandtia, p. 5. 6 CAMPODES. Campodes C. L. Koch, 1847. As has been pointed out in the paper on the Craspedosomatidae, this genus belongs to the lulidae. European writers have adopted Bollman’s identification of Cryptotrichus with Campodes. Silvestri goes so far as to make a family Campodidae, although this name is preoccupied in the Thysanura. Cryptotrichus is also preoccupied and has been replaced by Cleidogona. In a paper soon to ap- pear in the Amefican Naturalist a family Cleidogonidae has been recognized, to contain the American genera Cleidogona, Bactropus, and Pseudotremia. Doratonotus Pocock, 1894. This name has been used by Guenther for a genus of fishes, and was replaced by Doratodesmus. Haplosoma Verhoeff, 1893. This name is also preoccupied and was changed to Haplodesmus, the family to Haplodesmidse. Verhoeff has also recently published a genus of Geophilidae which he calls Haplogaster. Ihis name is at least twice preoccupied and may be replaced by Haplophilus. Macrotrichus Silvestri, 1896. With this name Silvestri proposes to replace Poratia Verhoeff, pre- occupied. The deficiency had already been supplied by Mastigona Ck., 1895. P^ROMOPUS Karsch, 1881. This genus is discarded by Silvestri. It is not only valid, but should be given family rank under the name Pseromopodidae. PlESTODESMUS Eucas, 1 849. This genus has been given as a synonym of Platydesmus Lucas, 1 843 , but the locality, description, and plates would seem to forbid an infer- ence of relationship with Polyzonium Brandt, to which Silvestri has reduced it. It is probably a valid genus. PSEUDoiUEUS Bollman, 1887. This genus is recognized as valid by Silvestri, although Bollman established it only as a subgenus and later withdrew it as having been based on an immature form *. Rhachis Saussure, 1859. Preoccupied; replaced by Rhacodesmus Ck., 1895; the only species is Rhacodesmus viridis (Saussure). Thrinciueus Porat, ,1895. The author of this genus takes Acanthiulus murrayi Poc. as his type, but applies the name to African species related to those described * Bull. 46 U, S. Nat. Mus. p, 138. HAPLOSOMIDJE. 7 by Karsch under Glyphiulus which according to Pocock is related to Trachyiulus Peters. Acanthoiulus is one of the Spirobolidse, and Thrinciulus must stand as a synonym, unless murrayi proves to be generically distinct from hlainvillei lycGuillou. At any rate, Thrinci- ulus will be a genus of Spirobolidae, while Porat’s Kamerun species are to be included in a genus Tophostreptus based on L. magnus (Karsch) ; I have compared the type with several specimens from Kamerun. The other species of Glyphiulus described by Karsch, G. scalatus, is the type of a related genus to be called Anastreptus. Trachysoma Attems, 1895. This name is preoccupied and has been replaced Trachygona ;• a family Trachygonidae has also been recognized. Verhckffia Broelemann, Dec. 1895. This name was given to supply the place of Tatzelia Verhoeff, but Haplogona Ck., proposed for the same purpose was already printed (Oct. 1895). The synonymy of some of the families recognized by Silverstri is as follows: Trachysomidae vSilvestri=Trachygonidae Ck. ; Canipodi- dae Silvestri=Cleidogonidae Ck. ; Trachyiulidae Silvestri, 1896= Cambalopsidae Ck., 1895 ; Nemasomidae Silvestri, i896=Isobatidoe Ck., 1895; Pyrgodesmidae Silvestri, 1896 = Sty lodesmidae Ck., 1895 (probably); Gervaisidae Silvestri, 1896 = Gervaisiidae Ck., 1895; Haplosomidae Silvestri, 1895 = Haplodesmidae Ck., 1895; Sphaerotheriidae C. T. Koch, i847=Zephroniidae Gray, 1842. Family CYCLODESMIDZB Silvestri, 1895. If we are to follow this author, his family must be reduced to Oniscodesmidae Saussure, i860, for Oniscodesmus is one of four genera which he includes in it. The others are Cyclodesmus, Cyrto- desmus, and Doratonotus (=Doratodesmus). After an examination of all four genera I am convinced that there is no family relationship unless it lies between Oniscodesmus and Cyclodesmus. On Cyrto- desmus and Doratodesmus the families Cyrtodesmidae and Doratodes- midae are established. Family HAPLOSOMID^ Silvestri, 1895. Under this name Sig. Silvestri includes a miscellaneous assemblage of all known diplopod genera with nineteen segments. Even Para- doxosoma Daday is among the number, although a family Paradoxo. somatidae was established six years in advance of the Haplosomidae. ♦Ann. N. Y. Acad. Sci. IX, p. 5, 1895. 8 CHORDEUMA TO IDEA. NOTH ON THE FAMIEIES OF CHORDEUMATOIDEA. A further examination of the relationships of the higher groups of Diplopoda has resulted in the separation of an order Coelocheta from the Merocheta and of an order Zygocheta from the Diplocheta. The Coelocheta will contain three suborders, the Lysiopetaloidea (Cal- lipodoidea), the Striarioidea (nov.), and the Chordeumatoidea ; the true lulidae and their allies comprise the Zygocheta. In the Chordeumatoidea several families may be distinguished as follows : Family CHORDEUMATiDiE C. E. Koch, 1847. Genera: Caseya, Chordeuma, Melogona, Mycogona, Under- woodia. Distribution : Europe and North America. Family CRASPEDOSOMATID^ Gray, 1842. Genera : Aporogona, Atractosoma, Attemsia, Basigona, Bomogona, Chelogona, Craspedosma, Grypogona, Haasea, Haplogona, lulogona, nov., type I. tyrolense (Atractosoma Verhoeff), Ochogona, Ornitho- gona, 710V., type O. helvetica (Atractosoma Verhoeff), Phanogona, Rhymogona, 7iov., type Rh. mo7itivaga (Atractosoma Verhoeff ). Distributioji : Europe. Family CEEIDOGONID^ Cook, 1896. Genera : Cleidogona, Bactropus, Pseudotremia. Distributio7i : Eastern North America. Family CONOTYEID^, nov. Genera: Conotyla, Trichopetalum, Zygonopus, Scoterpes. Distribzdion : Eastern North America. Family TRACHYGONID^ Cook, 1896. Genera : Rhiscosoma, Trachygona. Distribution : Europe. Family BRANNERIID^, nov. Genus : Branneria. Distribution : Eastern North America. Family HETEROCHORDEUMATID^ Silvestri, 1895. Genera : Heterochordeuma, Pocockia. Distribution : Malaysia. ♦American Naturalist, Dec. 1895, p. 1115. III. THE GENERA OF OXYDESMID.®. An analytical key to the genera of this family was published some months since but the study of more extensive and better material has led to the recognition of several new generic groups. The family has also been limited by the separation of some of the forms previously included, these being assigned to a new family called Pre- podesmidae t- The diagnostic characters of the genera now recog- nized are indicated in the following table, to which is appended a list of the species. A monograph of the family will appear in a few months. KEY TO THE GENERA OF OXYDESMIDA5. Dorsum densely beset with 4 to 6 transverse rows of coarse tubercles : Genus Scytodesmus Ck., Kamerun. Dorsum smooth, granular, or with three rows of polygonal areas, each with a large tubercle in the middle. ......... Submarginal ridge very oblique, broad and not prominent, remote from the margin ; pores in a broad, shallow’- depression in the middle of the ridge, hence also remote from the margin : Genus Plagiodesmus Ck., Congo Valley. Submarginal ridge nearly or quite longitudinal, parallel with and adjacent to the margin ; pores also not remote from the margin, located in the outer slope of the submarginal ridge Apical margin of last segment transverse and very broad, the lateral mar- gins parallel, convex, or divergent caudad ; apex proper also very broad, equal- ing in width the four setiferous tubercles, which are also located on the poste- rior margin and are not greatly exceeded by the apex : Genus Lacnodesmus, nov., type Z. campu Congo Valley. Apical margin of last segment more or less rounded or triangular in general outline, the lateral margins more or less converging caudad; apex proper dis- tinctly narrowed and exceeding the tubercles, one or both pairs of which appear to rise from the lateral margin : the posterior marginal tubercles at least projecting much farther caudad than the anterior, except in certain East African forms where the tubercles are greatly developed Dorsum smooth and shining ; under a lens finely rugulose or coriaceous ; no granules, tubercles or areas: Genus Mimodesmus Ck., East Africa. Dorsum with more or less conspicuous granules, tubercles, and areas. . . Segments with the tubercles not specially modified or increased in size ; copulatory legs bent at the middle of the last joint, and the apices inserted under the anterior edge of the large aperture : Genus Oxydesmus (H. & S.), West Africa. Segments 1-4. or some of them, with at least the median posterior tubercles hypertrophied ; copulatory legs extended and exposed *Proc. U. S. Nat. Mus. XVIII, p. 99, 1895. t American Naturalist, XXX, p. 415, 1896. Brandtia, p. 9. lO OXYDESMIDAl, Anterior segments dorsally somewhat thickened, raised and prominent over a bright-colored median area occupied by several mesial tubercles of each of the last two rows ; tubercles of anterior segments not conspicuously enlarged or coalesced into ridges or processes: Genus Lyodesmus, nov., tyv^ zoster, sp. 7i., East Africa. Some or all of the anterior segments (1-4) with the median or posterior tubercles conspicuously enlarged. ......... Segments 1-4 with the median and especially the posterior tubercles evi- dently hypertrophied, but not coalesced in the median line : Genus Rhodo- mus, nov., type Rh. priodus, sp. n.. East Africa. Segments 1-4, or some of them, with a conspicuous median process formed by the coalescence of two or more of the posterior tubercles. Third segment only with a process ; this is comb-like, very large, and re- curved so as to cover entirely the median part of the fourth segment : Genus Ctenodesmus, nou., type Ct. gibber, sp. n.. East Africa. Processes of first and second segments distinct, equal to or smaller than that of the third segment. . . . . . . . . . . . Segments 2-3 with the two median tubercles of the posterior row enormously hypertrophied into a long, slender, recurved, canaliculate process ; the first segment has also a small process : Genus Ceratodesmus, nov., type C. ajisatus, sp. n.. East Africa. Process not long, slender, and recurved; appearing rather in the form of a transverse crest of enlarged tubercles coalesced at base, and scarcely projecting beyond the posterior margin of the segment. ...... Fourth segment normal; copulatory legs slender, the two princinal rami extended into long, slender, falcate, approximate, incurved processes ; a slen- der ventral ramus extends mesad and distad to meet or cross its fellow : Genus Phobodesmus, nov., type Ph. cristatus, sp. n.. East Africa. Fourth segment with a distinct process similar to that of the third ; copula- tory legs pedicellate, stout, not deeply divided into slender rami : Genus Oro- desmus Ck., East Africa. EIST OF SPECIES. 1. ScYTODESMUS KRiBi Cook ; Kamerun ; Berlin Museum. 2. ScYTODESMUS coNNivENS, Sp. H. Smaller and more convex than Sc. kribt ; lateral carinse narrower, the edges with four or five sharp teeth ; copu- latory legs similar to those of Sc. kribi, but shorter and more robust ; length about 18 mm., width 5 mm. ; locality Bismarckburg, Togo Col- ony, collected by Conradt ; Berlin Museum. 3. Plagiodesmus obliques Cook ; Congo Valley ; British Museum. 4. Plagiodesmus occidentalis (Karsch) ; Quango; Berlin Museum. 5. Lacnodesmus campii (Cook); Congo Valley; U. S. Nat. Museum. 6. Lacnodesmus valgus, sp. n. Very distinct from the other species in size, dorsal sculpture, and form of copulatory legs. Segments with tubercles slightly developed, appearing as broadly rounded or flattened rugulose prominences or convex areas ; last segment with sides diverging, though the apex proper is rather narrow ; copulatory legs with the apical arms more crassate and the lateral spine longer than in L. campii ; length 70 mm. ; locality Africa ; Philadelphia Academy. OXYDESMIDA^. 1 1 7. Lacnodesmus FLABELLATUS (Cook) ; Congo Valley ; U. S. Nat, Museum. 8. Lacnodesmus ituri, n. Rather closely relating to Z. and of the same size and uniform dark color ; last segment with the sides scarcely diverging and apically narrower than in L. ca7npii ; segments 18-19 with their posterior corners narrower and more produced than in L campii ; locality Ituri Faehre, northwest of Albert Nyansa ; Stuhl- mann ; Berlin Museum. 9. Lacnodesmus thyriuotus, sp, n. Size and general appearance somewhat that of L. valgus ; differing in the greater convexity of the dorsum (pos- sibly a sexual character), the slight development of the dorsal tubercles and areas, the greater prominence of the submarginal ridge, the more rounded posterior corners of the carinse, less produced on posterior segments, and the shorter last segment with the two large marginal tubercles much less developed than in L. valgus ; color nearly black, a small, transversely rectangular light spot on each anterior subsegment, divided by a fine dark median line ; length about 70 mm. ; locality Jaunde Station, Kamerun ; Berlin Museum. 10. OxYDESMUS AFER Gray ; Senegambia ; British Museum. 11. OxYDESMUs ASABA, sp. H. Dorsum with tubercles distinct, though not prominent ; carinae with submarginal ridge moderately prominent, close to the margin, which is entire, or slightly sinuate on poriferous seg- ments ; pores opening obliquely outward, the surrounding depression very slight ; color of anterior subsegments nearly black, of posterior dull yellow with a longitudinal band of black on each side at the base of the Carina; length 45 mm., width 7 mm., the smallest species of the genus ; locality Asaba, River Niger ; British Museum. 12. OXYDESMUS BAROMBi, Sp. fi. Differs from O. johnstonei, sp. n. in the smaller size, the absence of the prominences of the vertex, the less con- vex segments, and the less prominent submarginal ridges ; color lighter than in O . johnstonei, dark coffee brown, sometimes lighter mesially; legs and antennse concolorous ; male 60 mm. by 10 mm., female 66 mm. by II mm. ; Barombi Station, Kamerun ; Berlin Museum. 13. OXYDESMUS CAFFEROiDES (Mattozo) ; Cabinda; Lisbon Museum. 14. OXYDESMUS DENTATUS, Sp. fi. Segments dorsally densely granulate, the three rows of tubercles prominent, more so than in any other species here referred to Oxydesmus ; carinse with edges distinctly sinuate-den- tate ; submarginal ridge smooth, prominent, broad ; pores removed from the margin by about four times the diameter of the rim ; color black, the submarginal ridge bright yellow ; antennae and ventral sur- face reddish; length about 60 mm., width 10.5 mm.; Africa; British Museum, a single female. 15. Oxydesmus dollfussii, sp, n. Segments with surface smooth and shin- ing, faintly and sparingly rugulose, divided into three rows of large areas with or without low tubercles ; carinae with submarginal ridge broad, especially on poriferous segments ; not sharp above, the pore in a depression of its highest part ; last segment with apex thickened, smooth and shining below, the apical setigerous punctations very close together ; color uniform, rather dark, dull, terra cotta red, tinged with pink below; length 85 mm. ; locality Africa ; British Museum. 16. Oxydesmus dorsalis (Murray); locality unknown. I 12 OXYDESMIDJ^. 17. OxYDESMUs FLAvoMARGiNATUs (Peters) ; “America?”; Berlin Museum. 18. OxYDESMUs FUAMBO, sp. fi. Segments with the surface granular-uneven, the areas rather distinct ; tubercles distinct, though slightly promi- nent, scarcely more so than in O. Icevis ; last segment with the apex squarely truncate, on each side three tubercles projecting from the strongly converging margins ; carinse with very entire and evenly rounded margins ; copulatory legs differing from those of all other Oxy- desmi in that the lamella is bent backward (downward) at a right angle ; the apex is then curved forward and has distally a deep rounded notch ; the lamella is not ’iserted under the edge of the aperature as in the other species, but lies on it, while the slender ramus is inserted as usual ; color black, shading into reddish ; the carinse bright yellow, and a dull yellowish median spot along the posterior margin of each segment ; length 45 mm,, width g mm. ; locality Fuambo, near Lake Tanjanika ; British Museum. 19. OxYDESMUs GRANULOSUS (Beauvois) ; Oware. 20. OxYDESMUS GRAYii (Newport) ; Sierra Leone : British Museum. 21. OxYDESMUs jOHNSTONEi, Sp . H. Vertex with two large prominences, the summits of which are hirsute with long hairs ; segments rugulose, the areas well defined, the tubercles large ; submarginal ridge sharp and prominent, close to the finely and widely serrulate edge ; pores in a deep depression in the middle of the submarginal ridge ; last segment rounded and not emarginate at apex ; copulatory legs without the prox- imal branch or process of O. grayzi ; color black ; length 80 mm,, width 12 mm, ; Rio del Rey ; British Museum. 22. OxYDESMUs L^:vis, Sp. 11. Similar to O. afer, but smaller and smoother. Segments dorsally indistinctly coriaceous, not granular, apparently smooth and shining ; tubercles obsolete ; carinse evenly rounded, the ridge moderately broad and high ; color nearly black, with the lateral half of each carina yellow, antennae and legs blight pinkish ; length of male 53 mm., width 9.5 mm, the anterior segments widest; Togo Colony, apparently common ; Berlin Museum. 23. OxYDESMUS LIBER Cook ; Liberia ; Berlin Museum. 24. OXYDESMUS MEDius Cook ; Liberia; Berlin Museum. 25. OxYDESMUS THOMSONii (Lucas) ; Liberia; type in Paris? 26. OXYDESMUS TOGOENsis Cook ; Togo Colony ; Berlin Museum. 27. OXYDESMUS TRicuspiDATUS (Peters) ; Guinea ; Berlin Museum. 28. OXYDESMUS TUBERCULiFRONs Porat ; Kamerun ; Stockholm Museum. 29. OXYDESMUS vALDAUi Porat ; Kamerun ; Stockholm Museum. 30. OXYDESMUS viTTATUs, Sp . 11. Segments with three rows of rather large tubercles ; snbmarginal ridge moderate, not sharp above ; lateral margin entire ; color dark brown, a median stripe of light coffee-and-milk brown, broader on posterior subsegments and occupying nearly half the width of the subsegment; length 70 mm., width ii mm. ; Cameroons and Old Calabar; British Museum. 31 OXYDESMUS XANTHOMELAS, sp. It. Evidently nearest related to (9, differing in the very slightly development of the low, rounded, tubercles and in the possession of a broad, rectangular, median yellow spot occu- pying the posterior half of that part of the dorsum : length of female 65 mm., width 11.5 mm, ; Kilmanjaro ; U. S, Nat. Museum. OXYDESMID^. 13 32. Lyodesmus zoster, sp. n. Segments dorsally with tubercles moderately prominent, especially those of the two posterior rows near the middle of the dorsum ; posterior margin on anterior segments somewhat raised in the middle, but no distinct process ; submarginal ridge prominent, rather broad, rounded and not sharp above ; pores located in a slight depression in the outer slope ; margins of carinee sinuate or subdentate ; copulatory legs subsimilar to those of Oxydesmus, forcipate, the broader mesial division deeply notched distally, the lateral division robust, in- curved : color probably brownish, the submarginal ridge and a median spot on the posterior margin of each segmen yellowish ; length about 50 mm., width 9 mm. ; locality Kawende, East Africa; Berlin Museum. 33. Lyodesmus FLAVOCARiNATUs (Silvestri) ; Juba River, East Africa; Genoa Museum. 34. Lyodesmus fischeri (Karsch) ; Massai Land ; Berlin Museum? 35. Lyodesmus effulgens (Karsch) ; Somali Land; Berlin Museum. 36. Rhododesmus PRiODUs, Sp. ft., Somewliat more slender and rougher than the other species ; anterior segments with median tubercles much en- larged, sharply conic ; dorsal surface beset with very distinct granules ar- ranged in clusters about the tubercles, and sometimes almost equaling these in size-; lateral carinse very long, the angles sharp and with four or five distinct teeth between; color dark reddish brown ; length 36 mm., width 5.8 mm. ; Dar es Salaam ; Berlin Museum. 37. Rhododesmus UNICOLOR (Cook) ; Mombassa ; Berlin Museum. 38. Rhododesmus mastophorus (Gerstsecker) ; Mombassa ; Berlin Museum. 39. Ceratodesmus ansatus, sp. n. Somewhat similar to the preceding species, but the dorsal sculpture less pronounced ; first segment with the two mid- dle tubercles of the last row coalesced into a distinct, though small pro- cess; second and third segment with a long, narrow, recurved process; carinse with angles sharp, the lateral margin irregularly dentate ; poste- rior margin with numerous fine, sharp teeth ; last'segment subtriangular, the sides distinctly converging, the marginal tubercles well developed ; color brownish, the carinse reddish, the dorsal processes bright red ; length of female 44 mm., width 7.5 mm. ; Tanga, Usambara ; Berlin Museum. 40. Ctenodesmus gibber, sp. n. Smaller and more slender than Ct. pectina- tus ; third segment with a more erect and somewhat narrower process composed of four equally hypertrophied tubercles ; margins of carinse with two large teeth between the angled corners ; last segment with apex narrow, produced ; color uniform reddish black, including the legs and antennse ; length 38 mm., width 5.5 mm. ; Jombene Range, East Africa; U. S. Nat. Museum. 41. Ctenodesmus pectinatus (Karsch) ; Wito, East Africa; Berlin Museum. 42. Phoeodesmus cristatus, sp. n. Second and third segments with a crest- like process rising from the posterior margin ; fourth segment quite normal, the median tubercles of the last row scarcely enlarged ; lateral margins with 6-8 small, irregular teeth ; dorsum nearly flat, the carinae horizontal ; color nearly black, the carinae becoming yellowish laterad ; length of male 45 mm., width 8.4 mm. ; Usambara; Berlin Museum. 43. Orodesmus FORCEPS Cook ; East Africa ; British Museum. 44. Orodesmus bicolor Cook; Tana River, East Africa; U. S. Nat. Museum. H OXYDESMID^. 45. Orodesmus ellipticus, sp. n. Size and general shape of Phobodes7nus cristatus, but markedly differing in that the terminal segments are gradually narrower, giving a long-elliptic outline. First segment with two median tubercles of the last row enlarged and coalesced ; segments 2-4 with the middle four tubercles of the last row and two of the second row completely coalesced into a large prominence ; other tubercles small, the surface of the segments not very densely beset with fine granules, and not rugulose as in other species ; lateral margins slightly and irreglarly sinuate or faintly dentate ; submarginal ridge not very strong, close to the edge ; color dark red, the processes of the anterior segments, the median third of the posterior part of the posterior subsegments, and the carinse, yellowish ; submarginal ridge, antennae, legs, and ventral .surface, red; Ngatana, vicinity of Kilmanjaro; British Museum. 46. Orodesmus camelus, sp. 71. Apparently closely related to O. bicolor, but differing in the somewhat larger size, greater dorsal convexity, and in the fact that while the processes of the anterior segments are some- what more developed than those of bicolor, the two middle tubercles are not nearly as large proportionally ; colors slightly darker than in bicolor ; length about 43 mm. ; width 8 mm. ; Tanga, Usambara; Berlin Museum. 47. Mimodesmus PARALLELUs Cook ; Karewia, East Africa ; Berlin Museum. May 30, 1896. IV. ON THE XYODESMIDi®, A NEW FAMILY. Under this name may be arranged several genera the pertinent affinities of which have thus far not been indicated. For example, Dr. C. O. von Porat has identified Polydesmus erythropus Lucas* from Kamerun, referring it first to Oxydesmust and then to Parades- mus+. My opinion is that it belongs to neither of these genera, and that the Kamerun species is not that described from Liberia by Lucas. It is proposed to name this new form, and to found on it a new genus, the binomial to be Scaptodes7nus porati. There is in the Ber- lin Museum a specimen from Kamerun belonging to an evidently related, though distinct genus. The dorsum is strongly and evenly convex, and densely and finely granular ; the lateral and posterior margins of the segments are distinctly and closely serrate-dentate ; the pores are borne on a small, distinct, lateral callus on the usual segments. For this the name Thymodesmus is proposed ; the type species, Th. pulvmar^ is known only from a female specimen with 19 segments ; it measures 23 mm. by 5.25 mm. These genera, together with Diaphorodesmus Silv., Cryptoporus Prt., and Xyodesmus Ck., constitute the African division of the family. There is some re- resemblance to the Oxydesmidae in the broad anal segment, but even this is, if closely observed, very different from that of any known member of the Oxydesmidae. The location of the pores is also very different, while the special development of the tubercles of some of the anterior segments is a character shared only by the Oxydesmidae. In South America are three or four genera in all probability more related to each other and to the above African forms than to other American families. I have, for instance, before me the types of Trachelodesmus and Thymodesmus, and the relationships, while by no means close, appear to be substantial. There are two other South American genera of Xyodesmidae. Hypodesmus is nearly related to Trachelodesmus, and is founded on Trachelodesmus constrictus (Ptrs). The other is established for Rhachidomorpha alutaeea (Ptrs). To * Lucas, Myriapodes du Gabon, Thomson’s Archives Entomologiques II, p. 442, 1858. “Patrie; cote de Malaguette.” This is the “Grain” or “Pepper Coast,” a name formerly applied to Liberia. t Porat, Bih. K. Sv. Vet.-Ak. Handl, XVIII, No. 7, p, 22, 1893. tibid. XX, No. 5, p. 34, 1894. Porat here proposes to refer even such species as Rhododesmus mastophorus (Gerst.) to Paradesmus, with the result of bringing members of at least three families into an invalid genus, the type of which belongs to a fourth family, and the name of which is preoccupied. Branutia, p. 15. i6 XYODESMIDAl. this series Peridontodesmus Silv., founded on Scytonotus woodianus (H. & S.) may probably be added. It will be distinguished from the American, and indeed from all the forms mentioned above, by the coarsely dentate segments ; but as nothing is known with regard to the location of the pores, it could not well be included in the synopsis. Neither this genus nor Trachelodesmus has any appa- rent relationship with Scytonotus C. L. Koch, such affinity having apparently been inferred without an examination of the types. KEY TO THE GENERA OF XYODESMIDvE. Segments 2-3 or 2-4 each with a pair of tubercles hypertrophied into long horn-like processes: Genus Diaphorodesmus (Silv.), type D. dorsicorms (Prt.), Kamerun. Segments with tubercles subequal. ........ Last segment broad and lamellar, basally emarginate on the sides so as to ap- pear constricted ; no distinct tubercle or conic process on the sides of the apical portion of the segment: Genus Cryptoporus (Prt.), type CV. verrucosus (Prt.), Kamerun. Last segment with a more or less narrowed truncate apex, and with a more or less distinct marginal tooth or process on each side of the apex proper ; in cases where this tooth is small the sides of the apical portion distinctly con- verge. .............. Pores located under the apices of the small carinae of segments 5, 9, ii, 14, 16, facing laterad; sternum of last pair of legs with two long spines ; segments 19: Genus Batodesmus, nov., type B. alutaceus (Ptrs.) ; Bogota. Pores, if present, with the usual formula, 5, 7, 9, 10, 12, 13, 15-19; carime either not narrowed or only the first three or four ; sterna either without spines or those at the base of the la.st legs are not larger than those on the middle segments. ........... Repugnational pores located in distinct excavations in the lateral margins of the carinae ; dorsum very slightly convex ; last segment with the marginal processes very strongly developed : Genus Scaptodesmus type Sc apt 0- desmus porati, sp. n. {Parades77ius erythropus Prt., non Lucas), Kamerun. Repugn atorial pores, if evident, located in a distinct marginal tubercle or callus, or the dorsum is strongly convex and the last segment has the marginal tubercles slightly developed. All the carinae broad and moderately long ; pores evident, located in a dis- tinct, marginal, bead-like callus: Genus Thymodesmus, 7iov., type Th. pulvi- 7iar, sp. n., Kamerun. Either the anterior carinae are evidently shortened, sometimes into spine-like processes, or the pores are not located in a marginal callus. Segments densely beset with sharply conic granules ; the edges of the rather broad horizontal carinae sharply dentate with somewhat larger cones ; anterior segments without enlarged tubercles; last segment with marginal processes well developed; pores obscure; sterna not spined : Genus Xyodesmus Ck., type X. pla7ius Ck., Togo. Segments above beset with rounded granules or tubercles ; carinae very narrow ; anterior segments shortened, with a transverse row of large tuber- XYODESMIDA^. 17 cles ; last segment with marginal processes inconspicuous ; pores located on the under side of the carinse ; sterna of the posterior pair of legs of each seg- ment with a long, incurved spine. ........ Lateral edge of carinse with a few distinct, rounded tubercles, on the infe_ rior face of the largest of which the pores are located : Genus Hypode sinus, nov., type H, constrictus Bogota. Lateral edge of carinse finely granulate or denticulate, the pore located under the most projecting part, but not on a special and distinct tubercle or callus • Genus Trachelodesmus (Ptrs.), type Tr. arcticollis (Ptrs.), Caracas. In all the genera of this family the sterna are unusually wide, but the American forms are further remarkable in that the sterna are spined. The absence of such spines characterizes the African sub- family Xyodesmini. The Trachelodesmini are characterized by the robust body, the short and pointed anterior carinae, the normal pore- formula, and the uniform development of the sternal spines, while Batodesmini have the body slender, all the carinse narrow and pointed, the remarkable pore-formula, and the development of very long spines on the sterna of posterior segments. The absence of pores from the seventh segment and their presence on the eleventh and fourteenth is a most remarkable character, and in connection with the flattened dorsum, elevated carinse, and nineteen segments, may easily necessitate the recognition of a monotypic family Batodesmidse, if forms connecting with Trachelodesmus are not discovered. The typi- cal and only specimen is dry and had been broken and mended with glue. It is not impossible that a segment had been lost, which would modify the latter part of the pore-formula, but as the specimen is a male the want of pores on the seventh segment can be certified. Mav 30, 1896. ■ I ■} V. CRYPTODESMUS AND ITS ALLIES. Under the name Cryptodesmus a great variety of forms has been described, many it is believed with no close affinities to each other or to the type of the genus. Much of this confusion has resulted from the fact that the characters of the typical species have thus far re- mained in doubt, and that the author of the generic name, without giving a satisfactory account of his type, described other unrelated species under Cryptodesmus, from which the characters of the genus have been wrongly inferred. To further complicate matters, Cryp- todesmus alatus (Ptrs.) was not a single species, but the three typical specimens belong to three species and two genera, both very distinct from Cryptodesmus. It is here proposed to indicate and briefly char- acterize these and other new genera. Family CRYPTODESMID^ (Karsch), 1879. Mitth. Muench. Ent. Ver. p. 143. The character ‘ ‘ pores wanting ’ ’ seems to apply nowhere in this family ; I have found pores in all the African, American, and Asiatic forms examined. It certainly does not apply to the name in the re- stricted sense in which it is here used. The genus Cryptodesmus (Ptrs.) is as yet monotypic, and was based on Cr. olfersii (Bdt.), a Brazilian species in the Berlin Museum. The antennae are distinctly clavate ; the first segment widely exceeds the head, and has the anterior edge even, but with a regular row of flattish granules just behind the margin all around ; it is as wide as the second segment. Segments dorsally ornamented with three regular, transverse rows of small, though distinct, subconic granules, each provided with a hair ; the lat- eral and posterior margins are sinuate-dentate. Pores of the usual dis- tribution, distinct, submarginal, located near the middle of the lateral edge on anterior segments, more remote and farther back on posterior. The dorsal surface has none of the flattened, radiating areas and stri- ations, nor the wide carinae of the African and other forms which I have proposed to call Pterodesmidae. The relationships of the other Neotropical genera do not seem to lie, however, directly with Ptero- desmus, but rather with African forms which have two or four longitudinal rows of tubercles enlarged and coalesced to form dorsal crests or carinae, and which under the following name it is proposed to separate from the Stylodesmidae. Brandtia, p. 19. 20 HERCODESMIDAl. Family HBRCODESMID^, nov. These agree with the West Indian forms in that the basal joint of the copulatory leg is inflated to contain and conceal the apical struct- ures, but differ in the presence of dorsal processes or crests, in the loca- tion of the pore at the posterior corner of the carina on a special lobe or tubercle, and in having the last segment greatly reduced and con- cealed in the penultimate. The dorsal surface, and especially the carinae, show in the American genera radiating lines and areas which in the Hercodesmidse have disappeared, except a slight trace on the carinae. The West Indian forms discussed above it is proposed to arrange in a separate family accompanied, how^ever, by a single African genus, Choridesmus Ck., known only from Liberia. Family CHYTODESMID^, ?iov. In all these genera the pore is located at the base of the posterior lobe or area of the carina ; the pore is thus remote from the margin, in the middle or somewdiat toward the posterior corner of the carina. The antennae, while distinctly clavate, are long and slender in com- parison with those of the Pterodesmidae. In all Chytodesmidae there is a more or less defined median depression or sulcus which does not appear in the Pterodesmidae. The type of the new genus Chytodes- mus is Ch. laqueatus (Ksh.), a Cuban species in the Berlin Museum. The dorsum is strongly and evenly convex, the carinae being narrow and depressed in the direction of the dorsal arch. Segments ornament- ed with three transverse rows of convex, smooth and shining areas which are rounded cr polygonal in shape, only those of the carinae showing the radial form. Pores large and distinct ; all poriferous carinae have four marginal areas, the others three. Genus Stictodesmus, nov. The type is St. creper, sp. n., from Bogota, in the Berlin Museum ; it is the smallest of the three “ types ” of Cry ptodesmus alatus (Ptrs). Much smaller and less convex than Chytodesmus, and with the dorsal ornamentation consisting of a few poorly defined large areas. The head is not entirely concealed by the first segment. Pores with a rather large orifice opening laterad just in front of the last marginal sulcus of the carina ; all poriferous carinae with four lobes, the others with three ; the first notch or sinus of the posterior margin on each side is very deep. Length 4.2 mm. ; width about .8 mm. Genus Docodksmus Cook, 1896. Brandtia, p. 5. Type D. vincentii (Poc.), from St. Vincent, West Indies, in the British Museum. The form of the segments and great width of the APORODESMUS. 21 horizontal carinae suggest the Pterodesmidae, but the presence of four longitudinal rows of distinctly more prominent areas or granules renders diagnosis easy. Although Mr. Pocock desired to call this genus Aporodesmus, the pores are present as stated under the family, but are difficult of determination on account of the fact that the whole dorsal surface is covered with a fine bloom or villosity. All poriferous carinae are four-lobed except those of the fifth segment. Last segment truncate at apex and with a distinct rounded lobe on each side. Genus Tridesmus, nov. Type Tr. sectilis, sp. n., from Porto Rico in the Berlin Museum. Similar in size and shape to the preceding, but with the dorsal sculp- ture less defined ; the poriferous segments are three-lobed like the others, to the fifteenth. Segments 15-19 are obscurely four-lobed. The posterior area of poriferous carinae is much enlarged and some- what produced, instead of rounded as in Docodesmus. Last segment very small, triangular, entire, rounded at apex, and scarcely exceed- ing the produced carinae of the next preceding ; the sinus of the nineteenth segment is also much nairower than in Docodesmus. Family PTERODESMID^ Cook, 1896. American Naturalist XXX, p, 417. All the members of this family are distinguished by the possession of very broad, nearly entire, radiately impressed carinae, and by the location of the pores in the anterior part of the carinae, sometimes near its anterior edge. The antennae are short and strongly clavate. Genus Aporodesmus Porat, 1894. Bih. K. Sv. Vet-Ak. Handl. XX, No. 5, p. 41. To this Porat refers three species, probably representing as many genera. The author bases his genus on Polydesmus gabonicus Lucas, but does not seem to have examined the type, without which precau- tion identification is mere conjecture, for Lucas’ descriotion will apply to many of the Pterodesmidae. If the view be taken that the type of Aporodesmus is not necessarily gabonicus, but rather the form which Porat calls falcatus, it may prove necessary to place Compsodesmus Ck. as a synonym of Aporodesmus, but as the copulatory legs, pores and other important characters of falcatus are still unknown, such a reduction might easily prove a mistake. In the meantime two forms which Porat has distinguished and figured under gabonicus may be safely laken up as good species under the names suggested, Aporo- desmus falcatus , sp. n., and A. subrectangulus , sp. n. According to Porat tlie seventh joint of the antennae is much shorter than the sixth, and in the African ‘ ‘ Cryptodesmidae ’ ’ the fifth joint of the antenae 22 APORODESMUS. is twice as long as the sixth. Neither of these statements will apply to any of the Kamerun Pterodesmidse known to me. Porat also describes the coxae as produced, but his figures show that the sterna are intended. *The Kamerun Pterodesmidae of the Berlin Museum are as follows : COMPSODESMUS PERLATUS Cook ; Jaunde Station. The larger of two new female specimens from the same locality. Measures 23.5 mm. by 8 mm. CoMPsoDESMUs LiMACiNus, sp. H. Strongly and uniformly convex, the carinae decurved in the direction of the dorsal arch ; all carinae somewhat fal- cate, with the posterior corners sharp and produced caudad ; sterna sparsely granulate, deeply impressed ; on posterior segments the sternum of the posterior pair of legs bears two conic processes, distinct on the sixteenth segment, long and devaricate on the seventeenth, small, pointed and approximate on the eighteenth. Color in alcohol horn brown, moderately light; length 33 mm., width 10 mm.; Jaunde Station, Coll. Zenker, a single female specimen. This is the largest and broadest known member of the Pterodesmidse, though C. perlatus is broader in proportion to its length. The size and more convex dorsum render C. limacinus strikingly distinct from all related forms. COMPSODESMUS CUPULIFER, Sp. H. Dorsum flat, scarcely convex in the middle, the carinae horizontal ; anterior segments with posterior corner scarcely produced ; sterna scarcely granular or pilose, without conic processes ; copulatory legs much as in C. pulcher Ck., short and simple, with a large distal cup-like excavation ; color dark brown above, whitish below ; length 26.5 mm., width 7.5 mm. ; locality Barombi Station. COMPSODESMUS SPINATUS, Sp. 71. Dorsum nearly flat, the median convexity greater than in cupulifer, less and narrower than in perlatus, with which there is agreement in outline and in the strongly falcate carinte; the ra- diating and other dorsal areas are somewhat more convex than in perla- tus, but not so heavily pigmented as in that species; posterior sterna (13-18) distinctly spined, the spines of segments i6 and 17 largest, the spines generally somewhat better developed than in perlatus, in which the same segments bear spines ; preanal scale with setiferous tubercles shorter than in perlatus; length 18.5 mm., width 6.5 mm.; Barombi Station. The copulatory legs are much as in perlatus, with the distal excavation deeper and the posterior spine slightly less developed. COMPSODESMUS KUAKO Sp . 71. Perhaps doubtfully referable to the present genus ; differing from the other species in the more slender form and somevvdiat deficient dorsal sculpture. The copulatory legs are very sim- ilar to those of the last species, but lack the spine at the posterior corner of the apical excavation. Dorsum somewhat convex in the middle, the carinae distinctly ascending so that their margins are as high as the middle of the dorsum ; three rows of convex areas on the convex median portion ; carinae smooth except for the radiating marginal im- pressions ; pores distinct ; sterna narrow, very minutely roughened and pilose, scarcely produced caudad. Color translucent whitish in alcohol, apparently unpigmented and probably white in life, as in Gypsodesmus; length about i8 mm., width 4.5 mm, ; "'‘Kuako bis Ki77ipoko,''' Congo, a single male specimen collected by Buettner. This is the only member of the Pterodesmidae yet known from the Congo Valley. OP HR YDESMUS. 23 In addition to the six African genera Aporodesmus, Pterodesmus, Lampodesmus, Gypsodesmus, Compsodesmus, and Tanydesmus there may be referred to this family the following from South America. Genus Chonodesmus, nov. The type is Ck. alatus (Ptrs.), one of the three specimens on which his Cryptodesmus alatus was founded. The remaining specimen of the three it is proposed to recognize provisionally as a second species of Chonodesmus, under the name Chonodesmus regularise sp. n. This differs from Ch. alatus in the more equal size and more regular distribution of the dorsal elevated areas, each of which bears a hair of considerable length. The carinae are less horizontal, so that the dorsum appears more convex, with the median arch less abruptly prominent than in Ch. alatus. The pores are located in exactly the same way, that is, in front of the middle of the carina, somewhat re- mote from the margin, opening laterad on the side of a small granule. There is a large series of Indian and Malayan forms referable to the Pterodesmidse. Mr. Pocock has already defined a few genera as distinct from Cryptodesmus, but the bulk of his descriptions stand under that name. In a small collection of Javan Diplopoda which came into my hands about a year ago are some specimens in all prob- ability at least congeneric with others which have been described from the East under Cryptodesmus. These Oriental Pterodesmidse have the dorsal tuberculation somewhat more pronounced than the African, and the radiating areas are less evident ; the antennae are somewhat more slender, and the copulatory legs are constructed on a slightly different plan. The pores are located at about the middle of the carinae, farther back than is usual in the African forms. Genus Pocodksmus, nov. The type is P. greeni (Poc.),"^ which seems generically distinct from Ophrydesmus in the hairy body, the short, thick antennae, the trituberculate preanal scale. The small size and apparently narrow carinae also indicate a distinct form. Whether the other Ceylon species is congeneric can hardly be inferred from Mr. Pocock’s description. In neither species is the location of the pores stated. Genus Ophrydesmus, nov. The antennae have the fifth and sixth joints greatly swollen on the outer side as they lie bent ; the secondary sexual characters con- sist in having the male legs finely hairy on the ventral surface, the sterna very deeply impressed both longitudianally and transversely, with the resulting four prominences very finely and distinctly granu- lar and pilose with short hairs ; on the sixth segment these sternal * Journ. Bombay Nat. Hist. Soc., 1892, p. 24. 24 CR YP TODESMOIDES. prominences are pointed ; the third joint of the third legs is enor- mously enlarged, and bears on its ventral face a large excavation surrounded by a fringe of hairs Genus Cryptodesmoides Pocock, 1895. Aun. Mus. Civ. Genova (2) XIV, p. 790. In a Burman species called Cr. fees Mr. Pocock has discovered pores located in the posterior half of the carinae, bat distant from the coiner. The plate does not agree with the description which says, “ well removed from and about equidistant from the lateral and poste- rior borders. ’ ’ According’ to the figure they are close to the lateral edge. Following the description we may suppose that the location is not so very distant from that in Ophrydesmus, but the genera are doubtless distinct by other characters. The Cryptodesmi of Weber' s Reise may probably be referred to Ophrydesmus. There appears to be another disagreement between the figures and the description t; according to the last the dorsal granules become less pronounced laterad, which is true of nearly all related forms, but according to the figures there is a median granular area, with the carinae nearly smooth. Family OTODESMID^, nov. Under this head are to be accommodated Trichopeltis Poc. and Otodesmus, a new genus founded on Trichopeltis watsoni Poc. The pores are located at the base of very broad, decurved carinse, near the anterior margin. Otodesmus differs from Trichopeltis in the shape * Ophrydesmus cede, sp. n., is the type; the carinae are horizontal, the color is uniform dark brown above, with pale legs and yellowish antennae ; length 20 mm., width 5 mm. ; locality the volcano Gede, Western Java. Beside O. gede, I have the following species : O. SCAURUS, sp. n. Dorsum more convex and more decidedly granular, the carinae narrower and somewhat more pointed than in the first species ; color lighter brown, the carinae yellowish ; antennae yellowish, more slender than in O. gede. Length 15 mm., width 3.5 mm. ; locality as for O. gede. O. TENGGER, Sp. fi. Differing from O. gede apparently mostly in size ; lon- gitudinal impression of the sterna broader and not so deep, and the four divisions not so prominent; length 13 mm., with 3.5 mm. ; Tengger, East Java, 1200 feet. O. PUGNUS, sp. n. Smaller and more slender than any of the above. In the pronounced granulation, somewhat narrower carinae and more convex dorsum resembling O. scaurus rather than the other species ; length 10 mm., width 2.2 mm. ; Pengalengang. The copulatory legs afford other diagnostic characters, but they are so com- plicated that it seems idle to attempt to describe them without the aid of drawings. t Pocock, Myriapoda of Weber’s Reise III, pp. 374-376, Pf XXII, figs. 13-15. INODESMUS. 25 of the posterior marginal lobes, in the dorsal sculpture, and in the copula tory legs, which are in Trichopeltis said to be so large and so deeply inserted that the legs of the sixth and seventh segments are widely separated to accommodate them. Otodesmus watsoni is from upper Burma, while Tfichopeltis bicolor is from Sumatra, and has the anterior and and poriferous (?) carinae dark, the others yellowish. Family STIODESMID^, nov. Returning to the African and' West Indian Cryptodesmoidae, we find genera which suggest certain members of the Hercodesmidae» but differ distinctly in that the last segment is not reduced and con- cealed in the nineteenth. Such are Stiodesmus Ck., from Liberia, and Cynedesmus Ck., from Grand Canary. In the last genus it seems to be possible to include Cryptodesmus ornamentatus Ksh., from Cuba. A specimen from Florida is more slender than Cynedes- mus, and the pores occur only on segments 5, 7, 9, 10, 12, 13 and 15. This has been given the name Psochodesmus crescentis * . The Javan genera Pronodesmus and Myxodesmus may also be placed here, although in habit they bear a very striking resemblance to such Hercodesmidae as Udodesmus and its allies. Lophodesmus Poc. doubtless also belongs in the Stiodesmidae ; at least it seems to resemble Stiodesmus and Cynedesmus more than do Pronodesmus and Myxodesmus. Family COMODKSMID^ Cook, 1896. American Naturalist XXX, p. 415, In this case the relationship to Cryptodesmus is doubtful, but the location of the pores in the anterior part of the segments is at least a suggestion of such afiinity, and the structure of the copula tory legs is somewhat similar to that of the Pterodesmidse. From a cave in Jamaica I have specimens of a genus related to Comodesmus. Genus Inodesmus, nov. Differing from Comodesmus in the somewhat more slender, monil- iform body, obsolete caring, more projecting last segment, and normal pore-formula, the pores located in shallow depressions in the lateral middle of the segments, not in front of the middle as in Comodesmus. The only known species, /. jamaicensis , is about equal in size to Como- desmus lanatus, and is lighter brown in color, but may be faded. * Psochodesmus CRESCENTIS, sp. n. First segment with horizontal rim very narrow ; dorsal granules less pronounced than in Cynedesmus ; carinse narrower; color light brown ; length 4.5 mm., width ,6 mm.; locality Crescent City, Florida. Collected by Mr. H. G. Hubbard, of the U. S. Department of Agriculture. 26 T HEL YDESMID^. Family THELYDKSMID^, nov. The genus Thelydesmus Ck. , while perhaps not unrelated both to Pterodesmus and to Comodesmus is so widely different that the assumption of affinity inside family lines seems unwarranted. The hispid and densely granular segments, and the comparatively narrow, tapering and dentate carinse, render it distinct from the Pterodesmidae, while the development of carinse, the location of the pores, and the structure of the copulatory legs, separate it from the Comodesmidse ; indeed it is not impossible that the discovery of other related forms may connect it with the true Cryptodesmidae, rather than with the other families now known from Airica. Family DALODKbMlD^, 7iov. No representatives of any of the above families are as yet known from East Africa, but a new genus from Central Madagascar is in the Berlin Museum. In general appearance it bears very little resem- blance to any of the families previously enumerated, and suggests rather Rhachidomorpha or Rhaccdesmus. A suspicion that this new" form might be related to Thelydesmus was destroyed by the fact that the fifth and sixth antennal joints are wider than long in Thelydes- mus, while in Dalodesmus {nov.) the antennae are decidedly slender and scarcely incrassate, so that the fifth and sixth joints are much longer than wdde ; veitex not granular ; first segment not concealing and scarcely wider than the head ; second segment distinctly, though not greatly, wider than the first ; dorsal surface of segments minutely and regularly punctulate, appearing nearly smooth, coveted with rather large and distinct rounded or subconic tubercles, those of the carinae and margins smaller and sharper ; carinae in width scarcely equal to half the body cylinder, inserted nearly on a plane with the middle of the dorsum, subhorizontal; anterior corner wanting, pos- terior corner acute, somewhat elevated ; pores large, located near the lateral margin, somewhat in front of the corner of the carina, facing laterad and dorsad ; legs rather long, especially the last joint ; sterna wdth a distinct cruciform impression ; copulatory legs very long, slender, and straight, extending farther ahead than the insertion of the sixth pair ; apex with numerous spiniform processes. The typi- cal and only known species is Dalodesmus tectus. It is uniform dark brown in color, and measures about 21 mm. by 2.8 mm. Family HYNIDESMID^, 7iov. Body minute and slender, clothed above wdth long, fine hairs. Dorsum very convex, the carinse inserted far down, and decurved so as to reach below the level of the ventral surface. Antennae short, vSubgeniculate and strongly clavate. The genus Hynidesmus {7iov.) has DORA TODESMIDAE. 27 the first segment semi-elliptic, pointed and unilobate at the ends, scarcely concealing the head in front, as wide as the head or the other segments ; second segment with the carinae much enlarged, subfalcate, pointed below, not concealing the head or first segment ; segments dorsally very finely roughened, but with no tubercles or granules, the hairs numerous and long, but not dense. Carinae at the posterior corner apparently composed of two lobes, the first of which projects very slightly backward ; the outer does not project caudad, but is extended cephalad (!) into a strong, pointed, somewhat depressed process which, when the animal is coiled, fits into a socket in the de- ficient posterior corner of the preceding segment. Pores located on the outer lobe of the carinae, at the sinus between it and the basal lobe; pore formula 5, 7, 9, 10, 12, 13, 15-18. Last segment normal, rounded, without carinae or tubercles, the apex narrow, somewhat projecting and decurved. Legs short and slender, inserted rather close together. The type is Hynides7nus lanifer , sp. n., pale brownish in color, nearly 5 mm. long and .75 mm. wide, with 19 segments. It was collected at Goenceng Filoe, West Java, at an altitude of 8,000 ft. At fust I supposed this might have affinity with the another minute, hairy, Javan form described as Helodesmus, but the characters men- tioned seem to be the only points of resemblance. A relationship of Hynidesmus with Doratodesmus seems more probable, notwithstand- ing the great differences in form and size. An immature specimen from Tengger, East Java, (2.000 ft.) is apparently congeneric with the above ; it has 17 segments and measures 3 mm. by .4 mm. Family DORATODESMID^ Cook, 1896. Brandtia, p. 7. The typical and first described species came from Java. Three specimens now before me seem to belong to two species, both appa- rently distinct, at least in size and color, from D. armatus Poc. Both species have the supplementary margin regularly pectinate. * Doratodesmus muralis, sp. n. Carinae and segments covered with large flat areas which are light colored, whilephe space between them is dark ; dorsal processes yellowish, nearly smooth ; anterior subsegments light grayish, except a v-shaped black spot in the middle of the dorsum ; carinae composed of three lobes ; length about 12 mm., width 2.6 mm, ; two specimens from Western Java, at an altitude of 4,000 ft. Doratodesmus vestitus sp. n. The areas which cover the segments are large, prominent, and with a conic tubercle in the middle : posterior segments with three transverse rows of tubercles, while only two rows appear in D. muralis ; dorsal processes also tuberculate in the same manner as the segments ; the whole animal is covered with a dense bloom which becomes light yellow as the specimen dries ; in alcohol the color is uniform light brownish ; length about 14 mm., width 3 mm. ; Gede, Western Java, 9,000 ft. 28 C YR TODESMIDA£. Family CYRTODESMID^ Cook, 1896. Brandtia, p. 7. Types of Cyrtodesmus veluiinus and C, granosus are in the British Museum. The latter species differs from the former in having the second segment much larger, the surface not velvety nor hairy at all, but beset with coarse granules ; anal valves entirely plane. It is proposed to separate this from Cyrtodesmus under the name Oncodes- mus granosus (Gerv.) Cyrtodesmus asper Ptrs. in the Berlin Museum appears to be more nearly related to C. velutinus. Family ONISCODKSMID^ Saussure, i860. The type of Oniscodesmus is also in the British Museum. It bears a very evident relationship to the Pterodesmidae. Oiiiscodesmus rubri- ceps Ptrs., in the Berlin Museum, differs in being more convex, and in having the pores borne on distinct and prominent tubercles. It ap- pears to be the type of a new genus and may take the name Lignydes^ mus 7ub7iceps (Ptrs.). For Oniscodesmus auraniiacus Ptrs. the name Detodesmus is proposed. It differs from both Oniscodesmus and Lignydesmus in the narrow areolate border of the segments ; from Eignydesmus it is further distinct in wanting a poriferous tubercle, and from Oniscodesmus in the broad sinus of the nineteenth segment and the larger last segment. Detodesmus aurantiaeus is light-colored and has the surface of the segments smooth and clean, while in the other genera they are covered with a black, soot-like powder. Family CYCLODESMID^ Silvestri, 1895. Ann. Mus. Civ. Genova (2) XIV, p, 747. As recently pointed out, this family if used as proposed by Sig. Silvestri is a synonym of the preceding. However, it does not ap- pear that the relationship of Cyclodesmus with Oniscodesmus is at all close, and it is probably better to retain both families. From a “small damp cave, Mandeville, Jamaica,” come several specimens considerably more robust than C, porcellayius Poc., and without the notch in the posterior margin of the segments. These it is proposed to name C. hubbardii, for Mr. H. G. Hubbard of the U. S. Dept, of Agriculture, who collected this and other interesting species of Diplopoda in Jamaica and Florida. The largest specimen of Cyclodesmus hubbardii 10 mm. by 2 mm. The surface of the segments is smooth and shining, but does not become white even when dried. On comparing the specimens with the types of porcellanus it appeared that the exoskeleton was more fragile, possibly owing to the subterranean habits of this species. Junk 10, 1896. VI. AN ENUMERATION OF THE PAUROPODA. An opportunity of comparing European and American species of Eurypauropus has directed attention to the fact that the structural di- veisities of the various members of this class have not received proper systematic recognition. The analogy of the Diplopoda shows that important and long-standing divergencies may exist with slight change in the antennae and mouth-parts, and that the structure of the exoskeleton is subject to little variation. This general stability is correlated with the similarity in the habits of all Progoneata, with the result that structural differences are of much greater morphologic and systematic importance than would be inferred while adherence is given solely to Hexapod analogies. In accordance with this view it is proposed to recognize among the Pauropoda three orders represented by Pauropus, Eurypauropus and Brachypauropus, and named respect- ively Cinona, Eepona, and Monona. The Cinona and Eepona have seven dorsal plates, four of which are double and cover two pedifer- ous segments each, while in the Monona each pair of legs has a dorsal plate to correspond. It may be objected that strict priority would compel the use of the name Pauropoda instead of Cinona, in which case Bollnian’s name Monopoda, though hardly appropriate, might be used for the class, but this change would probably not meet with general acceptance. The genera of Pauropoda are not numerous, and the described rep- resentatives of the three orders are not so divergent as to require more than one family in each. The Pauropodidae contain two genera, Pauropus and Stylopauropus"^; the Eurypauropodidae four, Eurypauropus, Trachy pauropus, Cyphopauropus t and Acopauro- pusf ; the Brachypauropodidae are as yet monotypic, known only from s Brachypauropus hamiger. That the Pauropoda are worthy of the class rank given them by Pocock can scarcely be doubted. The antennae alone would go far toward such a determination when we consider the stability of form * A new genus based on Stylopauropus atomus, sp. n,, and probably also including St. pedunculatus (Lubbock). t Established for Trachypauropus 7nargaritaceus Toemoesvary on account of the characters noted in the appended list of species. t Based on Eurypauropus ornatus Latzel, of which I have studied speci- mens belonging to the Berlin Museum. Brandtia d. 2g. 30 PA UROPODA. presented by the antennae of Diplopoda. Even Polyxenus has typi- cal diplopod antennae, and those of the related genera Eophoproctus and Saroxenus* are still more closely similar to the normal type. As a class the Pauropoda are one of three coordinates under the branch or subphylum Progoneata, the others being the Diplopoda and Symphyla, perhaps more nearly related to each other than either is to the Pauropoda. It seems strange that the evident and striking diplo- pod homologies of the Symphyla have been so long ignored. The sclerites of the head, for instance, are exactly those indicated by a comparative study of the sutures which appear in the various diplo- pod orders, while the anal stylets which have seemed so remarkably thysanuran, may easily be the homologues of those of the diplopod orders Monocheta and Coelocheta, which are again homologous to the anal bristles of the Merocheta. The coordinates of the Progoneata are the Opisthcgoneata, contain- ing the Hexapoda and Chilopcda. The contradictory opinions drawn from the morphologic researches of recent years give increasing strength to the suspicion that homologies outside of these lines can- not be safely maintained, and that a common origin of the Opisthog- oneata and Progoneata with the Arachnida, Crustacea and Malacopo- da is as remote as the annelid series, so that arthropodous connecting links may never have existed. It will accordingly be necessary to subdivide the group Arthropoda, with the result that if a relation- ship of the Progoneata with the Opisthcgoneata be allowed, there will be at least three phyla or subkingdoms. The first and perhaps the most primitive is the Malacopoda, to contain the Peripatidae. The second consists of the Progoneata and Opisthcgoneata, for which complex the name Eabrata is proposed. The third phylum Branchiata will probably not long suffice for the diverse forms which it is as yet supposed to contain, though the adoption of varied and parasitic habits in the Crustacea and Arachnida weakens the force of analogy in estimating the value of structural and developmental differences. THE SPECIES OF PAUROPODA. 1. Pauropus huxleyi Lubbock. Reported from various parts of Europe, England, Austria, Russia and Italy ; also from Massachusetts (Kenyon), and Pennsylvania (Ryder). 2. Pauropus lubbockii Packard. This is thought by Kenyon to be the same as the preceding ; but no adequate description has been published. 3. Pauropus impar, sp. n. Distinct in that the external flagellum of the stout branch of the antenna is not half as long as the other ; the glob- ulus is entirely sessile ; the last pair of tactile hairs are not as strongly * A new genus soon to be published in the American Naturalist. It is based on 5. scandens, a large Liberian species with long antennse and four large tufts of bristles to each segment. PAUROPODA. 31 developed as in the next species. The anal segment bears at apex above two rounded, apparently chitinous prominences ; below it has a slightly tridentate structure to which are articulated two fusiform processes differ- ing greatly from the figures of those of huxleyi as given by Lubbock and Kenyon ; length 1.3 mm., width about. 3 mm. Five specimens of this species were collected between Huntington and Cold Spring, Long Island, in May 1893. 4. Pauropus bollmani, sp. n. Specimens in the National Museum collected by Bollman, and doubtless the same as those reported by him from Bloomington have the dorsal hairs, especially the anterior, much more strongly clavate than those of the other species, or than the figures of Lubbock and Kenyon. There seems to be but one rounded and flattened superior knob on the anal segment, and the inferior processes are more slender and closer together than in P, impar, though the condition of the specimens is not such as to give confidence in this observation. The rami of the stout branch are about equal in length; larger specimens are about i mm. long. All are of a distinctly brownish tinge, but may have become stained from the rubber corks. 5. Pauropus filiformis, sp. n. Described from Austria by Latzel as P . huxleyi var. filiformis. According to Latzel the differences are con- stant, and as they are numerous and definable the reason for the varietal designation is not clear. 6. Stylopauropus atomus, sp. n. Differing from St. pedunculatus at least in the much smaller size, as mature individuals of both sexes are only .55 mm. long. The flagella of the antennse have a distinct, smooth, and somewhat enlarged base as in the species of Pauropus, while in pedunculatus, according to the diagrams of Lubbock and Kenyon (copy ?) the flagella are ringed nearly or quite to the base. The stout branch is nearly as long as the other, broader distad, and has the corner , cut away and the appendages attached as in the figures oi pedunculatus. Three mature specimens were found on rotting bark under fallen leaves in dry woods near Huntington in the present month. 7. Stylopauropus pedunculatus (Lubbock). Reported from England, Austria (Latzel), Germany (Haase), and Russia (Schmidt), but it is doubtful whether the continental species is the same as the English. 8. Eurypauropus spinosus Ryder. Known from Fairmount Park, Philadel- phia (Ryder), and from Bloomington, Indiana (Bollman). I have exam- ined Bollman’s specimens and cannot object to his determination, though certainty can hardly be reached without comparison with the types of spinosus or with material from the typical locality. I have found what seems to be spinosus in considerable quantity under bark of rotting poplar branches at Kirkville, Onondaga County, New York, on three occasions. The specimens differ from those from Indiana in being somewhat lighter colored and in having the dorsal hooks slightly less numerous. A large proportion of specimens taken in September were young, while in May no young were seen, g. Eurypauropus latzeli, sp. n. Without having seen American specimens of this genus Latzel has reported E. spinosus from Austro-Hungary. A careful study of Latzel’s description arouses the suspicion that his species is distinct, the following being the more important discrepan- cies : E. latzeh lives solitary under stones, has the segments strongly I 32 PAUROPODA. convex, all the sensory hairs jointed or feathered, and the legs project- ing somewhat beyond the margins of the segments. American spec - mens are gregarious and have always been found on or under rotting limbs; they are not strongly convex like Glomeridse ; the fourth and fifth pairs of sensory hairs are smooth, and the legs never project be- yond tho margins of the segments. Nearly all the other characters mentioned by Latzel might from their nature be supposed to be generic. 10. Eurypauropus cycliger Latzel. Known only from Austio-Hungary. Latzel’s description seems to indicate that the dorsal hooks are not pres- ent, and that there are many sensory hairs to each segment, which would point to generic distinctness frem spinosus, 11. Eurypauropus PociLLiFER Silvestri. Known only from Italy. The dor- sal spines are said to be in rows, which makes the generic designation somewhat doubtful, but whether the rows are longitudinal or horizontal is not stated. The dorsal tubercles are said to be "'in medio excavatisp but the meaning of this is not clear. The structures in question when viewed from above appear hollow in A. spmosus ; how they differ in E. pocillifer is not explained. 12. Trachypauropus GLOMERioiDEs Tcemocsvary. This has been reduced by Daday to a synonym of Eurypauropus cycliger Ltz., but if the charac- ters noticed under that species are really present at least the genus will will probably be maintained. 13. Cyphopauropus margaritaceus (Toemoesvary). According to the descrip- tion this species differs from the others in the following generic charac- ters ; Body elongate, three times as long as broad ; dorsum densely punctate and tuberculate, the margins with rough tubercles and a row of setse ; first segment triangular ; fifth segment (only?) sinuate at the sides; seventh segment sinuate at apex, near the anus with three long hairs: legs 6-7 jointed, the tarsus 2-3 articulate, with a single claw. Known only from eastern Hungary, and said to have the habit of a small species of Glomeris. 14. Acopauropus ornatus (Latzel). Specimens collected by Latzel were among Haase’s collection acquired by the Berlin Museum (No. 1645). They are generically distinct from Eurypauropus spinosus in that the dorsum is longitudinally carinate, the carime bordered by regular rows of fine tubercles and surmounted by a row of flattened hooks which ap- pear only on the carinae and lateral margins, where they are expanded into large, obovate, leaf-like, reticulate appendages; sixth and seventh segments not concealed, the sixth dorsall}^ unicarinate, the seventh pro- duced into a narrow apical process truncate at the end ; margins of seg- ments not notched; sensory hairs of fourth segment hot clavate. The dorsal parts are here much less strongly chitinized than in Eury- pauropus, and the habit is so different as to suggest the idea that were a larger series of forms known this genus would have been placed in a subfamily or family distinct from the Eurypauropodidse. 15. Brachypauropus hamiger Latzel. This most interesting creature measures .4 mm. by .2 mm, and is the smallest of the Progoneata ; it has a tiansverse row of four hamules on each segment, and two such rows on the first. Latzel found it but once, and its structure was so del- icate that drawings were not attempted. June 24, 1896. VII. THE GENERA OF ORYID^. This family is divisible into two subfamilies, the Oryini, having the spiraculiferous scutellum coalesced with its prescutellum, and the Orphnaeini, in which there is a distinct prescutellum usually larger than the somewhat triangular spiracle-bearing sclerite. In the Oryini the part in question is large and rectangular, as long as the tergites and conspicuous to the naked eye on account of the large size of the forms which compose this subfamily. The Oryini comprise two closely related genera, Orya, from the western part of the Mediterra- nean basin, and Parorya *, a new genus from Texas which may be distinguished from”air^merican members of the suborder by the characters mentioned. Moreover it is the most robust of American Geophiloidea. From Orya it differs in the much slighter development of the presclerites throughout the body, which is shown most strik- ingly in the anal (apodous) segment where the presternite is rudimen- tary in the female, while very distinct in Orya ; the spiracle-bearing scutella are also not as prominent in their anterior part as in Orya, where on anterior segments, especially, the lower corner in front is produced so that the spiracle appears to be located in a deep notch. The spiracles of Parorya are much longer and narrower than in the European genus. Both genera have the mouth-parts, anal legs and pleurae much the same, though more secondary sexual characters may appear when males of Parorya are examined ; the only feature of this kind to be noted at present is that in Parorya the apical joint of the genital palpi of the female, though small, is quite distinct, while in Orya it is reduced to a mere rudiment. The genera of Orphnseini are more numerous and somewhat more varied among themselves ; all are more slender and less depressed than the Oryini. No satisfactory examination of the mouth-parts and ventral pores has been made, but these are much alike in all the forms dissected, as should be the case in members of the same family. Diagnoses of these genera have been arranged as follows : Pairs of legs i6i or thereabouts ; spiraculiferous scutellum much smaller than its prescutellum ; two rows of long suprascutella, the lower broad, the upper narrow ; claw of maxillary palpus not pectinate ; frontal lamina not dis- * Based on g new’’ species with 113 segments, 120 mm. long when contracted in alcohol, and nearly 6 mm. broad : its locality is not definitely known, but it is probably from Texas, as it was among a collection made by Shufelt in Louisiana and Texas. Brandtia p. 33. 34 OR YIDyR, tinct; anal legs 6-jointed, long and slender, at least in the female; ventral pores on anterior segments in two transverse areas, a long posterior and a short anterior : Genus Heniorya nov.^ type H. lon^'issima, sj). n., Brazil. Pairs of legs iii or less, and with a different combination of the other char- acters mentioned. ............ One row of two distinct suprascutella to each segment ; pairs of legs 67-85 ; anal legs 6-jointed, straight, divaricate, slightly crassate in males ; Genus Orphnseus Mnt., type O, jyhos'phoreus (Linn.), cosmopolitan in the tropics. Two or three rows of suprascutella ; pairs of legs 93-1 ii. Spiraculiferous scutellum much smaller than its prescutellum ; the two sep- arated from the main tergite by three narrow scutella whose prescutella are very small ; anal legs slender, 6-jointed ; frontal suture obsolete ; claw of max- illary palpus with a fringe of strong pigmented spines; pairs of legs iii : Genus Ctenorya, nov., type Ct. jombene, si). ?/.. Jombene Range, East Africa. Spiraculiferous scutellum larger or scarcely smaller than its prescutellum ; anal legs 5-jointed, or if 6-jointed they are crassate and the frontal lamina is distinct, and the suprascutella are four, rather large, arranged in two rows ; pairs of legs 93-105. ........... Anal legs 6-jointed, incrassate; spiracles large, linear or semilunar; claw of maxillary palpus rather large, pectinate; sternites with a median longitudinal sulcus, the ventral pcyres arranged in four distinct, sub transverse areas ; Genus Aspidopleres (Prt.b.tvpe A. mterralati/s (Pvt.). Damaraland. Anal legs 5-jointed. ........... Frontal lamina coalesced ; tergites fovcolate ; claw of maxillary palpus ex- cavate, not pectinate ; Genus Notiphilides Ltz.. type A^. inaxiniiliani & S.). Mexico. .............. Frontal lamina distinct ; tergites quadriculcate ; claw of maxillary palpus pectinate: Genus Pentorva, nov., type P. af7'a, sp. 7i., {Ofi)h77CEUs \Aspido- ple7'es\fusatus Prt., TtoTi Koch,) Kamerun. The validity of the last subdivision must remain somewhat in doubt until the types can be subjected to more extensive study, as hatzel attempted only a generic sketch, and Porat thought to identify his animal with a previously described species. This last was re- duced by Meinert to a synonym of Ojya barba^ica, and Porat offers no reasons for setting aside this decision except that the description and plates would fit the Kamerun species if modified in two very impor- tant particulars. This is apparently an example of the reprehensible practice of re- ferring new species to old descriptions whose onl}^ applicability con- sists in the fact that they seldom touch upon characters of present utility in generic or specific definition. ■ July 21, 1896. GEOPHILOIDEA FROM LIBERIA AND TOGO, This group of Chilopoda is represented in Liberia by a few species which were named, described and figured over two years ago, but *^blication is still delayed, so that preliminary descriptions are offered here. The Liberian fauna is remarkable in that it seems to contain no species of the genus Geophilus, representatives of which have been described from all parts of the world, though many such reports are probably based on insufficient examinations. Thus in a recent paper on the Myriapoda of Kamerun Porat has described seven new species under Geophilus, of which few, if any, are true Geophili, and some of which differ from Geophilus by family characters. Of the several families and genera of Geophiloidea * those thus far collected in Liberia may be distinguished from each other by the following synopsis : Labial sternum divided ; prehensorial sternum very narrow, without chitin- ous lines ; anal pleuree inflated, with numerous pores ; sterna with a median y-shaped chitinous thickening ; Family Dicellophilidae. The genus Lamnonyx Ck. has the frontal lamina incompletely chitinized in front of the labrum, the margin of the labrum entire, and a claw-like callosity at the anterior corner of the cephalic lamina below. Labial sternum entire ; prehensorial sternum broad or with chitinous lines ; anal pleurae with few pores or eporose ; sterna without median chitinous thick- enings. . . . . . . . . .... Mandibles with a single pectinate and no dentate lamella ; labrum tripartite; claw of maxillary palpus simple : Family Geophilidae. The only Liberian genus is Schizotmnia Ck. which is distinct from Geophilus in the divided max- illary sternum, the distinct prosternal teeth and prominent claw of the prehen- sorial legs, the want of ventral pores and the compound or spinose claw of the anal legs. Mandibles with several lamellm, or at least with both pectinate and dentate ; labrum entire ; claw of maxillary palpus pectinately fringed. Ventral pores distributed in 2-4 indefinite areas located on the anterior and posterior parts of the sterna so that a central space is poreless ; a small sclerite above the spiraculiferous scutellum : Family Oryidse. The species which I have identified as OrphncBus phosphoreus (Linn.) is probably not indigenous in Lrrangement of the Geophilidse, Proc. U. S. Nat. Mus. XVIII, pp. ted ; ftrei^diagnosis are given in the appends ^ list of West African species. Brandt A p. 35. 36 GEOPHJLOJDEA. Liberia as it is never found except in European houses in wood perforated by- termites which are probably the food of this now cosmopolitan species. Ventral pores in a single median area ; no suprascutella. .... Ventral pores in a prominent, transversely oblong, chitinous plate located near the posterior margin of the sternum ; maxillary sternum divided ; anten- nae geniculate and clavate ; labrum slightly chitinized, edentate in the middle ; segments rugose and scabrous dorsally; Family Ballophilidae. The genus Ballophilus has the cephalic lamina and prehensorial sternum very broad, and the anal legs strongly crassate in both sexes, more especially in the males. Ventral pores in a depressed, circular, central area, with the reticulated in- tegument continued between the pores ; antennse filiform ; labrum well chitin- ized, strongly dentate medianly; segments smooth: Family Schendylidse- The Liberian genus Ctenophilus differs from Schendyla in the free labrum, three dentate lamellae, and slender legs; from Pectiniunguis in the distinct labial and maxillary sterna, the slender anal legs, and the free labrum ; from Escaryus in the presence of ventral pores, in the broad last sternum, in the two poriferous cavities of the anal pleurae and the deficient claw of the anal legs ; and from Nannophilus in the free labrum and the six-jointed slender anal legs. Of these new genera Ballophilus is the most remarkable. The unusuall}" large and clavate antennae and anal legs seem to be corre- lated with a peculiar habit of throwing first one end of the body and then the other when disturbed, instead of crawling like other Chilo- poda, this strange method of progression being swift and effective. Ballophilus is further protected by the unique color, often nearly black, and always inconspicuous in the twilight of the deep forests to which it seems to be confined ; at least I have never found it in coffee plan- tations or other open places, nor does it seem inclined to burrow in the ground, where Ctenophilus is commonly found. In two cases I have found a female Ctenophilus in a small cavity in buried rotten wood, coiled up with numerous young. A LIST OF WEST AFRICAN GEOPHILOIDEA. The Berlin Museum contains about 20 specimens of West African Geophiloidea, all from the Togo ColonyL These are very intere.sting in connection with the Liberian fauna. Some new species have been recognized, but they are so closely related to the Liberian that there can be no doubt that they fall within strict generic lines. The Kam- erun species reported by Porat are also introduced into this list, and a part of them are evidently congeneric with the Liberian. Thus while anything like careful collecting has been done at only three points along the enormous coast line, it appears certain that as in other parts of the world the Geophilidse, while not so widely distrib- uted as some larger Chilopoda, are much more extended than the Diplopoda. GEO PHJLO IDEA. 37 Family ORYID^. 1. Orphn^us phosphoreus (Linnseus). According to Pocock there is but one species of Orphnseus in spite of the several synonyms, and I am in- clined to this view after having examined material from the Malaysia, Aldabra Islands, Liberia, and "Rrazil. There is some variation in color- ing, but this is not constant, and depends somewhat on methods of preservation. Dried specimens frequently become purple, as was the case with the type of Linnseus. In Liberia it appears that Orphnseus lives only in civilized houses, and has doubtless been imported. It is moreover found only in the dry wood which has been riddled by term- ites, which probably form the chief food of this species, and would ac- cordingly be readily carried in shipping, which explains the remarkable phenomenon related by Linnseus. At Sierra Leone a specimen ap- peared quite as mysteriously upon the deck of our steamer. Moreover,' the other characters given by Linnseus do apply to many dried speci- mens, and if there is but one species of Orphnseus it may with consid- erable safety be called O. phosphoreus. 2. Pentorya afra Cook. In the preceding paper were given reasons for doubt of Porat’s determination of Himantarium. fusatum C. L. Koch from Kamerun, and the above new name was given rather than to con- tinue a false reference in the literature. IS Cook. Color in life fuliginous, or violaceous, or black, sometimes fading to yellowish in alcohol. Antennse with last joint equaling the two preceding ; basal lamina abqut three times as broad as long ; mandibles with one dentate and^.one (or two ?) pectinate lamellse, each many-toothed; sterna bisulcate^ roughened and hirsute; pleurss of last segment moderately enlarged and .pilose, the two large pores concealed under the hirsute last sternum which is broader than long (8:5); anal legs densely hirsute; pairs of legs 63-73. 'Phis species is very rare, and after much careful searching I succeeded fin bringing home only about a dozen specimens. In the Berlin Museum are several examples from Togo of a form somewhat more slender and lighter colored than B . clavicoruis, but I have found no characters sfficient to distinguish it specifically. 4. Ballophilus maculosus (Porat). There can be little doubt, that Porat’s Geophilus maculosus belongs in the present genus, and the presence of B. clavicornis at Togo suggests that the two species may be the same, but in that case Porat has overlooked the vef^: large pleural pores. Be- sides, none of the Liberian or Togo specimens has more than 73 seg- ments, while B. maculosus is c^dited w4th 77. Family SCHENDYIvID^. 5. Ctenophilus africanus Cook. Bright pink in life, pale yellowish brown in alcohol, the head darker. Prehensors slightly surpassing the frontal mar- gin of the head ; sternum longer than the coxa (3:2), narrowly emargi- nate anteriorly ; coxa wuth a rudimentary subapical tooth ; ventral 38 GEOPHILOIDEA. pores nearly in the middle of the sterna ; pleurae of last segment mod- erately enlarged, pilose, with the two large poriferous cavities partially concealed under the last sternum, which is hairy and about twice as long as broad ; pairs of legs 49-55 ; length 25-42 mm., width 1.5-2. 5 mm. The most common geophiloid in Liberian forests. A single specimen from Togo with 55 legs lacks the last pair and is in poor condition ; it may belong to this species. 6. Ctenophilus paurodus, sp. 71. Labrum with about 15 teeth (both Libe- rian species have about 30); claw of maxillary palpus very large, the fringe of spines very long ; prosternal teeth distinct, very broad ; coxae of prehensors without teeth ; the distal interior corner rounded ; ventral pores beginning on the second segment and extending about to the middle segments ; last sternum broad, the posterior margin notched in the middle ; two pairs of large pleural pores entirely concealed ; anal legs at apex with two slender spines as in Ct. africa7ius. Two speci- mens from Bismarckburg, Togo, collected by Buettner are in the Berlin Museum, a male with 43 legs and a female with 45 ; length of former about 14 mm. 7. Ctenophilus euentulus (Porat) ; from Kamerun, evidently belongs in the present genus ; it has 51 pairs of legs. 8. Ctenophilus simplex sp. 71. Differs from Ct. africanus in having a single process from the basal joint of the labial palpus, in the narrower and longer teeth of the median part of the labrum, the somewhat longer cephalic lamina and prehensors. Males of Ct. africaTins have 49-53 legs, while a male of simplex has 63 legs, a female 65. This species ap- pears to be extremely rare, only two specimens appearing in my collec- tion, while toward a hundred of africaTius were found. 9. Schizot.«;nia prognatha Cook. Body whitish, the head brown. Antennae with last joint exceeding the two preceding taken together ; cephalic lam- ina longer than broad (ii;8), somewhat broader anteriorly than posterior- ly ; basal lamina more than twice as broad as long (3:7), largely concealed by the cephalic and first scutum ; prehensors much surpassing the frontal margin, their sternum of nearly equal length and width, nearly twice as long as the coxa ; coxa obtusely bidentate ; pleurae of last seg- ment somewhat enlarged, with a few large and small pores concealed under the ultimate sternum which is broader than long (5:4); anal legs with a few very long hairs, the claw with several hairs or spines; pairs of legs 41-43 ; length 9 mm. ; width .5 mm. ; rare in Liberia. A speci- men in the Berlin Museum lacks antennae and anal legs, but from the remnant of its characters seems referable to this species. It has 43 legs. 10. Schizot/Enia quadrisulcata (Porat) from Kamerun has 45 legs and is 10 mm. long. It differs from 3". progyiatha at least in the possession of anal pores and in the unarmed anal legs. 11. ScHizoT/ENiA unguiculata (Porat); Kamerun. Prehensorial sternum an- teriorly bidenticulate, the coxa with a large fuscous tooth ; pleural pores 4-6, concealed ; anal pores wanting ; anal legs long, 6-jointed, the basal joints with short and long setas, the distal with a few long bristles ; pairs of legs 49. GEO PHILO IDEA. 39 12. ScHizoT/ENiA SUPPAR, sp. n. Closely allied to the last and perhaps identical with it, The prehensorial sternum has two very distinct teeth ; the coxa has a distinct rounded prominence on the anterior mar- gin somewhat above the middle, in addition to the strong distal dark brown tooth ; pleural pores 8-io, concealed. There are two specimens (one minus the head) in the Berlin Museum ; a male from Misahoehe, collected by Baumann, "Eus de7n Moder des Urwaldes ; a female from Bismarckburg has 51 legs, the last pair are more slender than in the male, nor are the basal joints setose. Both specimens are nearly white. 13. ScHizoT^NiA VARA, sp. H. Another near relative of the two preceding species, but apparently distinct in the very slight development of the large, tooth of the coxa of the prehensors ; the tooth of the claw is also much less pointed than in S. suppar. The prosternal teeth are even more distinct than in that species ; the pleural pores are only 3-4 ; the anal legs are very long and slender and nearly naked in both sexes ; and there are distinct anal pores, which are absent in the other species. Several specimens from rotting leaves and humus in the forests are more slender than S. suppar, and rather deep brown in color ; they reach 28 mm. in length and have 53-57 pairs of legs. 14. ScHizoT^NiA .^QUALis (Porat) ; Kamerun. Differs from the last at least in the absence of claws from the anal legs and the lack of anal pores ; legs 53 pairs. 15. ScHizoTAENiA POROSA (Porat) : Kamerun. The last pleurae have 4-6 con- cealed pores and others scattered over the surface, even on the sides ^ and dorsal surface ; anal pores present ; anal legs long and slender, un- armed ; pairs of legs 61-69, length 37 mm. ^6. ' Tretechthus ULiGiNOsus (Porat). After a careful consideration of the characters detailed by Porat it appears that his Geophilus uliginosus belongs to none of the genera known from Tropical Africa, and that it is not a Geophilus, hence the suggestion of the present new generic name. Family DICELFOPHIFID^. 17. Lamnonyx leonensis Cook. Body attenuate caudad, waxy yellow, the head dark brown. Prehensors exceeding the frontal margin ; sterna broad- er than long (6:5), longer than the coxae (3:2), slightly emarginate anteri- orly, and with an acute tooth on each side of the emargination ; coxa with two teeth; the two following joints and claw each with a similar trun- cate tooth ; legs hairy, the first pair about half as long as the second; anal pleurae with scattering pores of different sizes ; pairs of legs 49 ; length 20-35 mm. ; width 1-2 mm. This species was abundant at Sierra Leone, and a few specimens were taken in a garden at Monrovia, where it is probably not indigenous, as I never found it in or near the forests. 18. Lamnonyx togensis, sp. n. Nearly related to L. leonensis, but strikingly different in the want of pleural pores, the absence of which is correlated with a slighter development of the pleune and the proportionately greater size of the last sternum, which is more than half as long as the pleurae. This absence of pores is a remarkable character in the present genus, and I have accordingly searched for them with special care. 40 GEOPHILOIDEA, The specimen is mounted in balsam and has been examined under a high power, with the result that no apertures, tubes or glands can be made out, all of which parts are very easy of demonstration in similarly treated specimens of L. leonensis. Possibly as a compensation the anal pores and their glands are several times the size of those of L. leonensis. The length of body and number of legs are the same as in that species ; while the* armature of the prehensors and mouth-parts offer compara- tively unimportant differences. A single female in the Berlin Museum, from Bismarckburg, collected by Buettner. ly. Lamnonyx guineensis (Karsch), from the Island of Sao Thome, is said to have 51-57 pairs of legs. 20. Lamnonyx punctierons (Newport) is reported from Kamerun by Porat, and a specimen from Togo is perhaps referable to this widely distribu- ted species, July 21,, 1896. / IX. A SPINNING DIPLOPOD. It has long been known that females of Polydesmus and lulus lay their eggs in large clusters and protect them with walls of earth and vegetable matter, probably hardened by some secretion of the animal. There have also been numerous reports of the fact that for protection during moulting somewhat similar structures are provided. The latter, at least, are probably constructed of matter which has passed through the digestive canal, while the material of the former seems from the observations of Vom Rath not to have been thus elaborated. That observer is, however, certainly in error in supposing that the moulting-nests are not of definite and constant shape in the same genus or family. Our smaller North American Polydesmidae con- struct flattened or lenticular nests with thin, almost papery, walls, while Oxydesmi observed in Liberia have large ovoid chambers en- closed by thick clay walls. As in both cases the external and inter- nal appearance of the nests is characteristic and constant, and as the wall is complete and apparently of the same structure throughout, we are obliged to infer either that the material for finally closing the entrance is supplied in the manner suggested, or that the animals provide themselves in advance with enough extra clay or vegetable debris to finish their cell. It is certain also that the matter voided after the creature has taken up residence in its chamber is added to the walls, so that it becomes antecedently probable that the whole structure is built in the same manner. With regard to the Coelocheta (Lysiopetaloidea and Chordeumatoi- dea) a different condition is known to exist. Fanzago has reported for Lysiopetalum, and Waga and Latzel for Craspedosoma, the exist- ence of nests composed of webs, but of the character of either the nests or the webs no account is current in the literature of the sub- ject, so that it was not without surprise that recently I found on top of Lookout Mountain, Tennessee, what appear to be young specimens of Cleidogona (Cryptotrichus Packard), in loose, white, silky cocoons built on moist decaying leaves. The cocoons were so transparent that the form and color of the inclosed animal was not entirely ob- scured ; indeed I opened the first one to see what sort of larva looked so much like a young Cleidogona ! The occupant of the second co- coon soon became restive when his establishment was exposed to Brandtia, p. 41. 42 CLEIDOGONA. light and air, and made his escape by gnawing a small round hole in the side of his nest, or rather by Dushing aside the threads of the loose network. The whole structure is about 5 mm. across, while the animal it contained is about 9 mm. long by i mm. broad and has 26 segments. The threads of the network are of very even and con- stant diameter, slightly more than a thousandth of a millimetre. They are transparent and glassy in structure, and appear to be glued together where they lie one upon another. The open meshes are usually many times the diameter of the thread. The fabric is about equally loose throughout, the side which lies against the leaf not dif- fering from the other parts. That these cocoons were spun by their tenants does not admit of doubt, but the method of their production is not yet known. The Coelocheta and Monocheta (Stemmatoiulus) differ from all other Diplopoda in the possession of large papillae articulated to the apex of the last segment, and Vom Rath has found that these are connected with a complex internal apparatus. That these papillae are spinning- organs has naturally suggested itself, but how the spinning is accom- plished is still a mystery. The papillae in question are each tipped with a slender fine-pointed hair. This is hollow at base, but I have been unable to follow the lumen to the apex. The supposed spin- ning organs of Scolopendrella are also tipped with a hair, but the source of the thread seems not to have been ascertained. That the Coelocheta and Monocheta possess functional spinning- organs which are wanting in the other orders would demand recogni- tion in a natural arrangement of the group. This has already been accorded by their ordinal separation from the polydesmoid and iuloid forms, from which they differ by complexes of other characters. July 25, 1896. I X. AN AMERICAN GLOMEROID. Up to the present time no member of the order Oniscomorpha has been reported from any part of the Western Hemisphere. It was ac- cordingly a pleasant surprise to collect recently at Auburn, Alabama, numerous specimens of what appeared at first sight to be a small pill- bug, but which closer examination and dissection showed to be a new form evidently related to the European family Glomerididae, but dif- fering in characters which indicate long divergence from the same ancestors. The body is smaller and less heavily armored than in true Glomeridae, though the general shape of the sclerites is much the same. The last three pairs of legs of males are modified in nearly the same manner as in the Glomerididse and Gervaisiidae, showing that in the Oniscomorpha as in each of the other Diplopod orders, the copulatory legs are of more or less completely independent structure and origin, and may, indeed, have been the ground of ordinal differentiation. That the differentiation of copulatory legs was begun before the power to coil into a sphere was perfected is evidenced by the fact that Zephronia, Glomeris, Gervaisia, and Onomeris (nov.) each use a somewhat different method. Thus a coiled Glomeris has a groove in the edge of the posterior corner of the first segment into which the corners of the second and third segments are fitted, and the middle of the last segment rests against the first, which is thus partially exposed. In Onomeris there is no groove at the posterior corner of the first seg- ment, but a notch somewhat above connects with a curved groove or excavation not present in Glomeris, below and in front of which is a broad, striated margin which fits under the projecting corners of the other segments when the creature is coiled. Onomeris is able to coil somewhat more effectively than Glomeris, and the first segment is completely concealed. In Zephronia the antennae are located at the sides of the head and are accommodated under the projecting lobes of the second segment, while in Glomeris and Onomeris they are inserted near the middle of the face ; they are closer together in Onomeris, and strongly genicu- late, the sixth joint being folded against the third ; they lie in two deep excavations which extend from near the median line to the single vertical row of eyes. This is in strong distinction to Glomeris where the vertex is not hollowed out to accommodate the antennae, which project beyond the sides of the head into the space covered by the second segment. Brandtia, p. 43. 44 ONOMERIS. The gnathochilarium of Onomeris is similar to that of specimens of Glomeris eonspersa, which do not, however, resemble Latzel’s dia- gram very closely. The seventeenth pair of legs of males have the coxae much en- larged, but the other joints, three in number, are reduced to mere rudiments. The eighteenth pair of legs is five-jointed, with the coxae coalesced. The copulatory legs are, as stated, not greatly difierent in general shape from the European forms, but they are provided with large finely corrugated processes from the posterior face of the two last joints, in addition to processes from the median face similar to those of Glomeris. The last segment is broadly emarginate posteri- orly ; in size it is about equal proportionally to that of Glomeris. The surface of the segments is smooth and shining, but is finely punc- tate under sufficient magnification. The first segment has two trans- verse striations and the anterior part of the second five or six ; the anterior edge of the other segments is finely margined and trans- versely grooved ; the groove is finely and regularly crenulate like the suture of certain Polydesmoidea. In color the specimens are light grayish or horn-brown, darker along the posterior margin and laterad, though the posterior and lat- eral edges may be pale ; on some specimens there is a distinct lateral pale spot. The deficient pigmentation is no doubt correlated with the slighter chitinization of the American genus, though the small size is also to be considered. The largest specimens of Onomeris measure less than 5 mm. by about 2 mm.. When disturbed they coil at once into a ball and do not move for two or three minutes. Over a hundred specimens of both sexes were found among rotting leaves and pine needles, near Auburn, Alabama, during the present month. The species is to be called O. underwoodii for my friend Professor E. M. Underwood, whom I was visiting wEen Onomeris and other interesting Myriapoda were collected. To give this new genus a proper assignment in the classification is not easy. Evidences of affinity with Glomeris are evident and nu- merous, but many of the common characters are also shared with all the members of this peculiar order, and as noted above, the differen- ces, while perhaps not of great importance, are such as would indi- cate long separation from the European stock. A Javan Glomeroid in my collection is evidently much nearer to the European than is the American form, although the Malayan species are apparently distinct from the European by generic differences, so that I have for some time been intending to designate Glomeris infuscata Poc.^ as the type * Pocock, Myriapoda of Weber’s Reise, p. 324, PI. XIX, figs. lo-iob. Apioineris infuscata comes from Sumatra. GLOMEROJDEA. 45 of a new genus to be called Apiomeris. Even the European Glome- rididae have never been subjected to a careful comparative study of structural characters, and many species have been described with little more than colors to distinguish them. A comparison of the European material at my disposal makes it appear probable that the recognition of a separate family for Onomeris will be found warranted and desirable, and there appears to be in the literature no account of any form which could serve as a connecting link between Glomeris and the American genus. The chief objection to this position now apparent is that consistency may compel the elevation of Eatzel’s Glome?is minima (=Tatzelia Bollman, preoccupied, =Glomeridella Broelemann) to an equally exalted position. However, the great similarity in form of the eighteenth pair of male legs in forms so widely separated as Glomeris, Onomeris and Gervaisia, is good reason for supposing that it is a primitive character, and that the remarkable forcipate eighteenth legs of Glomeridella are indicative of equal or greater divergence. It is moreover similarly suggestive that the very numerous species of Glomeris, have, as far as known, copulatory legs and their accessory pairs built on the same general plan, the dif- ferences being so ."light that European writers have seldom taken the pains to point them out, a task which would doubtless give some in- teresting results in the line of constant minor differences. KKY TO THE FAMIEIES OF GEOMEROIDEA. Tergites eleven, strongly roughened, their posterior margins prominently elevated and tuberculate ; mentum entire ; Family Gervaisiidse. Tergites twelve, smooth and even ; mentum divided. ..... Eighteenth legs of male four-jointed, stout, forcipate ; nineteenth pair not exceeding the eighteenth, three-jointed ; their median lamina large and flab- ellate ; eleventh tergite much reduced: Family Glomeridellidae, nov. Eighteenth legs five-jointed, slender, not forcipate, exceeded by the very robust nineteenth pair ; median lamina small and rectangular or Ungulate ; eleventh tergite normal. .......... Seventeehth legs of male with coxa enlarged, otherwise atrophied to a three- jointed rudiment; second segment with a broad, striated anterior and lateral margin separated by a distinct transverse groove ; antenna accommodated in deep excavations in the vertex, which leave a narrow ridge around the margins and a prominent median crest: Family Onomerididse, nov. Seventeenth legs of male of the normal form, and only slightly reduced ; second segment smooth except for a marginal notch or groove at the inferior posterior corner ; antennae not included in fossae, their terminal joints extend- ing beyond the sides of the head : Family Glomerididae. July 29, 1896. XL THE LARVAE OF STEMM ATOIULUS. In February of the present year I collected some peculiar diplopod larvae on an island in the St. John’s River, Grand Bassa County, Li- beria. This island, though not in a swampy region, is covered with pandanus and other low-land trees, and is no doubt flooded in the rainy season. The larv^ were about 30 in number, and pure white in color, except a single reddish ocellus. The largest individuals measure nearly 5 mm. in length by about .8 mm. ; some are as much as one-third smaller. They were guarded by a large female of Stem- matoiuhis bellus, which was coiled about them in a manner stiongly suggesting that of Scolopendra or Geophilus. They were not en- closed in any nest or egg-case, but were lying on moist sand under a rotting log. That they were newly hatched is made certain by the fact that they had not yet begun to crawl about, and that some are still included in a membrane covering the entire animal without dis- tinction of parts, apparently the same motionless, pupiform condition in which the young of lulidse leave the egg, but which is not known in Polydesmidse and Glomerididse. On escaping from this mem- brane the larvae of Stemmatoiulus have five-jointed antenae and three pairs of six-jointed legs, but they differ from all known six-legged diplopod larvae in the possession of thirty-five segments. The six- legged larvae of Polyxenus have five segments, while those of the Glomeroid^, Polydesmidae, and lulidae have seven. The mother of these larv^ had 49 segments, and numerous adult specimens of the same species vary only between 48 and 50. In connection with the other remarkable and probably primitive characters of Stemmatoiulus this peculiarity of the larvae is very sug- gestive. The possession of this great number of segments increases the size of the embryos and limits their number ; other Diplopoda lay about three times as many eggs. We may hence infer that this is another of the primitive features of Stemmatoiulus, and that the many-segmented embryo and larva are ancestral conditions from which a majority of the Diplopoda have advanced. In the light of this fact the hitherto unique embryonic feature of the epimorphous Chilopoda, the possession of the full adult complement of segments and legs, has great importance, as confirming the view that the Prog- oneata and Opisthogoneata may be remotely descended from a many- Brandtia, p. 47. 48 ST EMM A TOIUL US. segmented ancestor, and that this apparently anomalous character of the Epimorpha is after all, merely a primitive condition which most Eabrata have left. Moreover, it no longer throws doubt upon the affinity of the Epimorpha with the other Chilopoda, or of these with the Hexapoda. That a many-segmented embryo was the condition which obtained among the remote ancestors of the Labrata may, then, be inferred from three considerations : 1 . A many-segmented condition not being an advantage would not be called into existence or favored by natural selection. 2. This condition appears in two unrelated orders of Labrata whose nearer relatives are distinctly anamorphous. 3. It also appears in the viviparous Malacopoda which are probably more nearly related to the ancestors of the Labrata than any other extant forms. The hexapodous condition of the young of Diplopoda furnishes ac- cordingly very slight argument for any direct or genetic relationship between the Diplopoda and the Hexapoda, as Bollinan declared, even without the support of the present evidence. That the embryos and larvae of the ancestors of Stemmatoiulus were at one time provided with numerous legs as well as segments is indicated by the existence of a nearly complete series of chitinized, flattened, lanceolate processes or propodia, more highly developed than the so-called ‘ ‘ buds ’ ’ which precede the legs of some other Di- plopoda, and doubtless homologous with the ventral appeadages of Campodea and other apterous hexapods. We have here additional reason to believe what has long been suspected, that these abdominal appendages are rudimentary legs. Moreover, the protrusible vesicles of Machilis and even the collophore of the Collembola may easily be remnants of the likewise protrusible nephridial structures which ap- pear on the coxse of all the legs of Colobognatha, and on some of those of Coelocheta, but have otherwise disappeared from the Diplo- poda. At any rate, these various problematical appendages in Cam- podea, Japyx, and Machilis, if ancestral, rudimentary and trans- formed from their original function, do not furnish reasons for suppos- ing their possessors to have any close affinities, nor break the force of the wonderful structural diversity long known to exist between the various Thysanuran families. Lubbock pointed out over twenty years ago that the association of Campodea and Japyx with the Thysanura proper (Lepisma, Machilis) must be considered as merely provisional, and subsequent discoveries have rendered their distinctness more and more apparent. The proposition of Grassi and Stuinmer-Traunfels to place Japyx and Campodea with the Collembola as a suborder En- ST EMMA TOIUL US. 49 tognatha does not give a more natural arrangement, which will hardly be reached until four orders of so-called Apterygogenea are recognized*. The recent refusal of certain eminent entomologists to to take this step because the groups are nearly monotypic, is, to say the least, unscientific, but not less so than the willingness to recognize as natural groups assemblages of apparently unrelated forms. Thus it has long been known that Machilis and Tepisma have more in com- mon with the Orthoptera than with Japyx, Campodea, and the Col- lembola, l ut each new work attempts the task of describing together the indiscribably diverse, with a variety of minor changes which go far toward rendering classification chaos. The Coelocheta are probably the nearest relatives of the Monocheta, and it is a suggestive fact that among them no six-legged larvae have been discovered, the youngest yet reported having 15 segments and the same number of legs, in a species referred by I^atzel to At- tractosoma. In Tysiopetalum forms with less than 30 segments are not yet known. It is also remarkable that the prolegs of the larva of Stemmatoiulus are all of the same form and size, so that it appears not unlikely that at the first moult nearly the full complement of legs for the thirty-five segments would have been added. In these young larvae of Stemmatoiulus the exoskeleton is still membranous and shrivels up on drying ; there appears to be no trace of the peculiar ornamentation of the surface, and the pleural sutures are indistinguishable. A peculiar feature is the presence on the ventral parts of each side of the tergites of four long bristles, of which the inferior is longest ; no other bristles can be made out, and it is prob- ably these same bristles which appear in the adult in a rudimentary condition. The expansion of the ventral part of the body necessary when the legs are added would bring the bristles more nearly into the subdorsal position which they occupy in the adult. The great size of these structures in the young larvae is another character tend- ing to show their primitive nature and probable homology with the enormous spines of geologic forms. The bristles of young Stemma- toiuli are, however, simple, slender, and smooth, in strong distinction to the barbed, clavate hair-structures of young Merocheta ; they are equal in length to nearly one-third of the diameter of the segments to which they are attached, and are directed obliquely backward, so that when a coiled larva is viewed from the side it appears to be sup- plied throughout with slender legs. The Colobognatha are another order presenting many primitive characters. No six-legged larvae are known. The youngest * Halida3^’s duplicate generic name for Japyx, Dicellura, would furnish an appropriate ordinal designation, while the name Rhabdura is proposed for an order to contain Campodea. 50 ST EMMA TOlUL US. specimens of Polyzonium germanicum reported by Latzel had 14 or 15 segments and 12-16 pairs of legs. At Sierra Leone in December 1893 January 1894 I collected numerous specimens of Siplion- otidse (P/iznotus africanus Ck.), a family even more primitive than the Polyzoniidae. The smallest of these has 15 segments and 21 pairs of legs, that is, it has only two terminal apodous seg- ments, as in adults. This is further evidence of the proposition that Diplopoda originally had numerous segments and legs, even in the larval condition, for lulidae and other forms in which the pleurae have disappeared always have several apodous segments when immature, a series of segments being added at each moult, while the legs for them are not supplied until the next July 30, 1896. XII. NEW AMERICAN PLATYRRHACID.^. The genus Platyrrhacus was based by C. L. Koch on a Brazilian species, Polydesmus scaber Perty, or at least on a specimen so deter- mined, and described as being slightly convex, densely granulate, and with a row of distinct, pearl-like tubercles along the posterior margin of each segment. There are said to be two other rows of somewhat smaller tubercles placed wider apart. Although the carinae are said to be strongly toothed, they appear from the plate that the teeth are broad and rounded. After studying the description in con- nection with that of another American species described by Koch, Platy7rhacus riifipes, the opinion has been gained that it would not be safe to identify it, even generically, with any of the material which has come into my hands for study. The genera Stenonia and Stosatea Gray, having had no types as- signed to them by their author, remained nomina nuda and can be neglected, as Gray himself did in the preparation of a list of the Myriapoda of the British Museum (1844), a paper overlooked or mis- placed in the bibliographies of Latzel, Daday and Silvestri. If we accept for Stenonia a type proposed by a later writer * it must be Polydesmus dentatus (Olivier), a species not known to Gray, a result certainly not in the interest of either justice or clearness. In the past no rule has been consistently followed, the name Stenonia having been applied to a great number and variety of genera, both American and Malayan. When sufficient types and new material have been studied, it will be necessary to rehabilitate the genus Platyrrhacus, which may supplant one of those proposed below, but Stenonia should be allowed to rest in oblivion. The number of American genera of Platyrrhacidse is doubtless very considerable, representatives of several of which have been loaned me by various museums. The impossibility of describing all these under one name, and the apparent artificiality of any arrange- ment which does not take into account a wide complex of the charac- ters of a series of related forms, are my excuse for proposing the new genera whose diagnoses are arranged in the following table : Dorsum strongly convex, the carinse strongly decurved in the direction of the dorsal arch; carinae of segment 19 distinctly pointed; last segment broadly *Gervais, Apteres IV, p. 95, 1847. Brandtia, p. 51. 52 FLATYRRHACID^. rounded ; sterna with a prominent conic process at the base of each leg ; dorsal surface finel}’- and evenly granular, the granules not prominent ; tubercles dis- •tinct but not conspicuous; areas indistinct ; lateral margin of carinae with sev- eral long, bead-like, smooth granules, which are somewhat pointed on posterior segments ; anterior and posterior margins of carinae finely and sharply den- tate; Genus Cyphorrhacus, 7tov., C. andhtus, sp. n., Ecuador, U. S. Na- tional Museum. Dorsum moderately convex or flat, the carinae more or less horizontal ^ carinae of segment iq broad and rounded at the posterior corners; last seg- ment usually broadly truncate ; geneia which may not be clearly eliminated by the above characters have either the sternal spines small or obsolete, the dorsum coarsely granular or finely so, and with the tubercles obsolete, the areas distinct or entirely absent, the lateral margins subentire, with coarse granules, or with large pointed teeth, the anterior margins of the carinae smooth or beaded with rounded granules, or some combination of these characters. ............. Carinae of moderate width, with a prominent, square shoulder at base in front; dorsum densely beset with distinct, prominent, rounded granules and divided into three transverse rows of polygonal areas, each of which has a dis- tinct, though not conspicuous tubercle in the middle ; tubercles of the poste- rior row much more conspicuous than the others ; first segment with an ante- rior row of large tubercles, behind which is a distinct transverse depression vertex densely and finely granular, prominent and subcri.state on each side of the sulcus ; pores removed from the margin by five or six times the diameter of the small ring ; clypeus smooth and shining below : Genus Psammodesmus nov., type Ps. cos, sp. n., Colombia, Philadelphia Academy. Carinae broad and wing-like, with the shoulder slight or wanting ; dorsum with very fine, sparse or flattened granules ; or if the granules are strong there is no trace of areas, and the carinm not very wide, the first segment has no large anterior row of granules, the vertex is rounded and with rather scat- tering coarse granules, or the pores are close to the margin or removed from it by only two or three times the width of the large and prominent rings. . Pores situated in the posterior half of the carina, just above the lateral edge, which is entire or subgranular ; dorsal surface divided into large areas sug- gesting those of Polydesmus; copulatory legs small, rather slender, the apical processes subequal in length, strongly divaricate: Genus Nanorrhacus nov,, type N. lucioe (Poc.), Santa Lucia, West Indies, British Museum. Pore near the middle of the carina, remote from the margin by at least two or three times the diameter of the ring; lateral edge nearly always distinctly granulate or dentate, with the posterior corner acute and distinctly produced , areas either wanting or numerous ; copulatory legs with the ultimate divisions unequal, approximate or connivent. ........ Dorsum densely beset with equal, prominent, rounded granules; the tuber- cles similar, but larger, the three rows about equally distinct ; areas entirely wanting; lateral margins with large, rounded, bead-like granules; first seg- ment with no specially enlarged granules in front ; copulatory legs divided immediately above the hairy portion into two slender subapproximate pro- cesses, the anterior of which is longer ; both processes are strongly curved cephalad, and the larger is bent nearly into a circle : Genus Rhyphodesmus, nov., type Rh. terininalis, sp. n., Museum. ) ( y ■