SMITHSONIAN MISCELLANEOUS COLLECTIONSVOLUME 106, NUMBER 18
ON THE EVOLUTIONARY SIGNIFICANCEOF THE PYCNOGONIDA(With One Plate)
BYjoi<:l w. hedgpethMarine Biologist, Texas Game, Fish and Oyster Commission
(PuiiLICATIUN 3866)
CITY OF WASHINGTONPUBLISHED BY THE SMITHSONIAN INSTITUTIONMARCH 24, 1947

SA'IITHSONIAN MISCELLANEOUS COLLECTIONSVOLUME 106, NUMBER 18
ON THE EVOLUTIONARY SIGNIFICANCEOF THE PYCNOGONIDA(With One Plate)
BYJOEL W. HEDGPETHMarine Biologist, Texas Game, Fish and Oyster Commission
(Publication 3866)
CITY OF WASHINGTONPUBLISHED BY THE SMITHSONIAN INSTITUTIONMARCH 24. 1947
BALTIMORE, MD., U. S. A.
ON THE EVOLUTIONARY SIGNIFICANCE OF THEPYCNOGONIDABy JOEL W. HEDGPETHMarine Biologist, Texas Game, Fish and Oyster Commission(With One Plate)INTRODUCTORY NOTEThe Pycnogonida, or sea spiders, are an anomalous class or sub-phylum of marine arthropods, unknown except by name to most zo-ologists. They are of no economic importance to man, and of littlediscernible significance in the natural order of things. Yet within thelast lo years more than 50 papers dealing with these creatures havebeen published, and the complete bibliography now comprises severalhundred titles. More than 500 species have been described, but thereare relatively few parts of the world in which the pycnogonid fauna isadequately known, and the actual number of extant species may beconsiderably larger.Also known as Pantopoda, the Pycnogonida have often been con-sidered an "appendix" to the Arachnida in comprehensive treatises,but they have no real relationship to the arachnids, since at no stagein their development do they have a cephalothorax or prominent abdo-men. Their relationship to the Crustacea is even more tenuous, forthey do not have biramous appendages, and their own peculiar larvalstage, the protonymphon, is distinct from all other arthropod larvae.They are characterized by an extreme reduction of the body, very longlegs, which house the sex glands and diverticulae of the gut, and a sub-sidiary pair of egg-bearing legs, or ovigers, which are present in allmales but lacking in the females of some genera. In addition to thewalking legs, which are usually 8 in number, but may occasionally be10 or 12, and the ovigers, there may be a pair of chelate appendages(cheli fores) and sensory palpi. The presence or absence of theseaccessory appendages constitutes the basis of classification within thegroup. There is a simple nervous system of ventral ganglia, and arudimentary circulatory system. There is no respiratory system or anyspecialized excretory organ, although in the males of many speciesthere is a specialized cement gland which is believed to be of use inattaching the eggs to the ovigers of the male, who carries them aroundSMITHSONIAN MISCELLANEOUS COLLECTIONS, VOL. 106, NO. 18
2 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. I06
until they are hatched. Shortly after hatching, the larvae of manyspecies become encysted in hydroids, sea anemones or small medusae,where they live as parasites for a time. Some species are found inholothurians and at least one species spends its early life in the mantlecavity of bivalve mollusks.Pycnogonids are found from the littoral zone to depths of morethan 2,000 fathoms, in all oceans. Some live on sargassum at thesurface w^hile others appear to be bathypelagic. They vary in size froma span of 3 mm. for small littoral species to about 50 cm. in some of theabyssal forms.In the course of my taxonomic studies of the collections of Pycno-gonida in the United States National Museum, the Museum of Com-parative Zoology, and the Peabody Museum at Yale, I have examinedseveral hundred specimens, including examples of many of the knownlo-legged forms, and in this paper I shall attempt to bring togetherin a coherent manner somiC of the speculations inspired by this largemass of material. Unfortunately, little is known about these creatures,especially the deep-sea forms, other than that which can be surmisedfrom their pickled corpses. The following discussion, therefore, isnot intended to blaze new paths in invertebrate morphology and evolu-tion, but simply to suggest some directions in which future investi-gators might profitably set out. In other w^ords, I have gatheredtogether some speculations on the possible causes of some observedefifects, for, as Aristotle said, "Nature does nothing which lackspurpose." ^I wish to acknowledge with thanks the generosity of Dr. ^^'^aldo L.Schmitt, Head Curator of Biology of the United States NationalMuseum, in loaning me material and literature, and I also wish tothank Dr. Isabella Gordon, of the British Museum, for her patientanswers to my persistent correspondence at a time when the BritishMuseum was in the front line of battle (and suffered accordingly)and for the gift of a specimen of Nymphon hiemale from the Dis-covery collections.
I. PHYLOGENY AND PATTERNS OF VARIATIONFrom time to time attempts have been made to divide the Pycno-gonida into orders, but the families are so closely related, and theirboundaries so broken down by transitional generic forms, that noneof these attempts have been successful. An attempt to separate thefamilies on the basis of the presence of ovigers in both sexes, or in the
1 Gen. Anim. II, v. (741 b, 4-5), Loeb Classics ed., p. 207.
NO. l8 PYCNOGONIDA—HEDGPETH
males only, for example, breaks down in the genus Pallenopsis, whichhas well-developed ovigers in the male and rudimentary ones in thefemale. Some workers have assigned this genus to the Phoxichilidi-idae, others consider it a member of the Pallenidae. Diagnoses basedon the retention or loss of chelifores in the adult cannot even beapplied at the generic rank, particularly in the predominantly achelategenus Achelia, which has several chelate species. Hence, when onecomes to draw a family tree, it looks more like a tangled web, ora bush with anastomosed branches, as in figure i.
DodecolopodaPentanymphoT) Decolopoda PentacolossendeisNymprion V--.?.,-- Colossendeis
AcMla -^f^7^^
PentapycnonPycnogonum Phoxichilidium
Tanystylum
Fig. I.—Hypothetical family tree of the Pycnogonida. (Transitional genera inbackhand lettering.)
It will be noted that this family tree—or bush—rises from a hemi-spherical base. This base is the quantitative diagram of the familiesrepresented in figure 2, together with diagnostic drawings of the mostcommon types of pycnogonids. Most of the forms illustrated in thisdiagram represent the type genera of the families concerned.Such attempts as this to erect a family tree naturally bring up theproblem of the roots, that is, the phylogenetic relationships of thePycnogonida with other arthropod groups. This is probably the mostdifficult problem connected with these animals and may never besolved. Snodgrass, in the most recent and lucid discussion of arthro-pod evolution (1938) is none too certain of the affinities of thePycnogonida, beyond placing them in the Chelicerata together withthe Arachnida and Xiphosurida. But, if we accept his restriction
SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. I06
Table 1.
—
Synopsis of the families of Pycnogonida
Family
NYMPHfemurfir6t tibia eye tuber
clielifore
palpus oviqer
.second tibia CO'Xji
tarsus
, propodus
AMMOTHLIDAL
lanystylum orhiculore
ac
TANYSTYLIDAE ^
Fk;. 2.—Dia

NYMPHONJDAL
TANYSTYLIDAL
PALLLNIDALFig. 2.—Diaograin ot qualitative and quantitative relationships of the Pycnogonida.

NO. 1
8
PYCNOGONIDA—HEDGPETHTable 1.
—
Synopsis of the families of Pycnogonida—Continued
Family
6 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL, I06than to any other group. But they are none too sure on the point, andconclude their discussion of affinities with this escape clause (p. 179) :
"Aber daruber ist das letzte Wort noch nicht gesprochen." Caimanand Gordon (1933) are more positive in their view of the arachnidaffinities of the Pycnogonida, and advance the theory that discrepan-cies between the Pycnogonida and the Arachnida might be accountedfor by metameric instability in the cephalic region of the Pycnogonida.It does not seem to be that this is a tenable view, for although there is
Fig. 3.—Some transitional forms of pycnogonids. a, Pigrogromitus timsanus(after Caiman, 1927) ; b, Paranymphon spinosum; c, Rhynchothorax mediter-raneus (after Dohrn, 1881) ; d, Decachela discata; d', tarsus and propodus; e,Pentacolossendeis reticulata. The line indicates i mm.
evidence of instability in the Pycnogonida, I am inclined to believe thatit is too recent to be of phylogenetic significance, and is of a differentcharacter than Caiman and Gordon apparently believed.However, I do agree with these authors that certain fossil formsfrom the Lower Devonian are not pycnogonids, and I further sus-pect that the group is a fairly recent one, and is still undergoing activeevolution.Certainly an inspection of the transitional and anomalous genera(see fig. 3) lends support to this view, for they present examples ofmissing links which would delight and confuse a paleontologist. Thisoccurrence of diversified forms connected by numerous transitional
NO. 1
8
PYCNOGONIDA—HEDGPETH 7
types suggests youth rather than age, for we would expect a loss oftransitional and experimental forms in an old group.There is, for example, the curious genus Pigrogromitus Caiman(1927) from the Suez Canal.- The body type of this genus is thesame as that of Pycnogonum, the type and only genus of the familyPycnogonidae, which is without chelifores and palpi and lacks ovigersin the female. But Pigrogromitus, with its lo-jointed ovigers in bothsexes and chelate chelifores, must be placed in the Pallenidae. Noris it the only transitional genus in this family ; it represents but oneof the extremes of variation.Most members of the Pallenidae are of the long-legged, extendedtype, but there are several compact disciform genera, such as Pseudo-pallene, which suggest affiliation with the Tanystylidae. Without itspeculiar propodus, which Hilton (1939a) considered to be a familycharacter, Decachela appears to be another transitional form betweenthe Tanystylidae and Pallenidae. The possession of a large spine on thesole, which is opposable to the terminal claw, the two forming a sub-chelate structure, cannot be considered a character entitling this formto family status, although it may be a variation sui generis. There arethe usual eight joints in the leg instead of the seven suggested byHilton's statement, "legs apparently seven jointed."The genus Pallenopsis, a transitional form between the Pallenidaeand the Phoxichilidiidae, has already been referred to. The phoxi-chilidiid characters of this genus are the possession of femoral cementglands in the male and the overhanging prolongation of the cephalicsegment. Because of its lo-jointed ovigers, present in both sexes butreduced in the female, I consider it a pallenid.Thus, in one family alone, there are transitional forms indicatingaffinities with three diverse groups considered worthy of family rank.If we consider the Tanystylidae, which is related to the Pallenidaethrough such genera as Pseiidopallene and Decachela, we find that itin turn is related to the Pycnogonidae through the genus Rhyncho-thorax, which has a body form approximating that of Pycnogonum.Inasmuch as the Tanystylidae are almost inseparable from the Am-motheidae, observations on the relationships between either of thesefamilies and the other families are mutually applicable. The principaldifferences between the two families are that the proboscis is usuallylarge and bulbous in the Ammotheidae and that the palpus has more
2 Hilton's (1942c) Pigrogromitus robtishis from Unalaska is actually Pycno-sonm strongyloccntroti Losina-Losinsky, and constitutes an extension of rangefor that species from its type locality (48° 58.2' N., 140° 35.3' E.) to Alaskanwaters.
8 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. Io6joints. The Ammotheidae (and Tanystylidae) comprise a diversegroup of forms whose common characters are the reduction of thechelifores (or, when chelate, the chelae are small and lack teeth onthe fingers), well-developed palpi, and ovigers in both sexes, usuallywith compound spines on the terminal segments.Among the interesting genera in the Ammotheidae are CilunculusLoman (1908), the male of which has femoral cement glands openingthrough a prominent dorsal tube as in Pallcnopsis and some speciesof Anoplodactylus (Phoxichilidiidae), and Paranymphon CauUery( 1896) , which has heretofore been referred to the Nymphonidae. InParanymphon the palpi are seven-jointed, whereas in Nymphon theyare always five-jointed, but the principal reason for removing thisgenus from the Nymphonidae is the discovery of a somewhat similarform, Ainigma Heifer (1938), in which the ammotheid affinities aremore clearly marked, and which in turn seems to be more closely re-lated to Paranymphon than to any other genus. These two curiousforms are very similar in body form, with high dorsal tubercles onthe well-separated lateral processes, and simple tarsal joints.In contrast to the diversity of form in the Ammotheidae, Tany-stylidae, Pallenidae, and Phoxichilidiidae, the families Nymphonidae,Colossendeidae, Pycnogonidae, and Endeidae are remarkably uniform.The Pycnogonidae and Endeidae are monogeneric families, and thereare but two indubitable genera in the Nymphonidae and Colossen-deidae, apart from the extra-legged forms, which are a special case.These latter families, however, possess large numbers of closely re-lated species.The most conspicuous thing about this pattern of variation is theway in which it is correlated with extra-legged forms. It will be noted,from an examination of figure 2, that in families where variation ismanifested at generic rank (as indicated by the high proportion ofgenera to species), lo-legged forms do not occur (so far as we know),whereas in those families in which variation is more active at thespecies rank, several lo-legged forms have appeared. The only ex-ception to this generalization is the Endeidae, a small monogenericfamily which may actually be an offshoot of the Pycnogonidae.It will be noted that I have not attempted to indicate the compara-tive ages of the various families or branches in my diagram. It isusually contended that the Pycnogonidae are the most specializedfamily, and the Nymphonidae the most generalized group, retainingmore of the primitive attributes than the Pycnogonidae. The Colos-sendeidae are intermediate, according to this view, and the otherfamilies branch out from the tree more or less according to individual
NO. l8 PYCNOGONroA HEDGPETH 9fancy or taste. When lo-legged forms were first discovered, it wassuggested that they were the primitive types, and should be at thebase of the tree (Cole, 1905), but this suggestion was made beforethese forms were well known. It now seems more reasonable to sug-gest that they are recent innovations in form, at least as far as thepresent pattern of variation is concerned. The occurrence of these 10-legged forms in three widely divergent branches suggests a commonorigin for these branches. For that matter, none of the families areactually different enough to enable one to assign any to a higher orlower place on a vertical scale, and the pattern of variation in thePycnogonida is not amenable to a lineal or two-dimensional inter-pretation, but is three-dimensional, the various families or branchesdiverging in all directions from a central or nuclear type.How such a structure might be bound into relationship with theother groups of the Arthropoda, is difficult to say. However, it isprobably no more difficult to visualize a relationship of three-dimen-sional structures than it is to decide just where, in a simple branchingpattern, the Pycnogonida stem out from the remaining Arthropoda.All phylogenetic trees and speculations are influenced by honest errorsin evaluating characters, but as Snodgrass said in his concludingremark (1938, p. 149) : "Every biologist must have a working creedof phylogeny, but he should not too implicitly believe its tenets."
II. TEN-LEGGED FORMSForty years ago, most zoologists who interested themselves in thematter believed that the Pycnogonida possessed but four pairs ofwalking legs, and considered that feature one of the diagnostic charac-ters of the group. Although the first lo-legged form had been col-lected by James Eights on a voyage to the South Shetlands in 1829and described, with an adequate figure, in 1835 under the name Deco-lopoda australis, it was generally ignored by naturalists.^ For ex-
3 For the melancholy story of Dr. James Eights (M. D.) and his Antarctictravels, see The Reincarnation of James Eights, Antarctic explorer, by JohnM. Clark (Sci. Month., vol. 2, No. 2, pp. 189-202, 1916), and James Eights, apioneer Antarctic naturalist, by W. T. Caiman (1937a). The latter paper isbased on the earlier one, but a bibliography of Eight's writings has been addedand the discussion of his zoological discoveries is more extended. There is alsoan interesting diagram of extra-legged pycnogonids, and a facsimile of theoriginal figure of Decolopoda australis. Unfortunately the color of the copy fromwhich the facsimile was made is poor; it should be red instead of slate brown.Further information on Eights and his contemporaries will be found in a briefpaper by Lawrence Martin, E^rly explorations and investigations in southernSouth America and adjacent Antarctic waters by mariners and scientists from
10 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL, I06
ample, Hoek, in his monograph on the Challenger Pycnogonida(1881, p. 6), dismissed Eight's Decolopoda with these words:
". . . . I have not been able to ascertain whether this is a good genus,nor where it has been found." Apparently he did not see Eights'paper. When the Rev. T. R. R. Stebbing wrote a series of populararticles on pycnogonids for Knowledge in 1902, he considered Deco-lopoda an amateurish blunder and denied that there could be such athing as a lo-legged pycnogonid. A few years later, Dr. J. C, C.Loman, a well-known Dutch zoologist and author of several paperson the Pycnogonida, published a paper to the effect that Decolopodamust have been a monstrosity. Hardly had the ink dried on thesecontributions when the South Polar expeditions began to return withnot one but two species of lo-legged pycnogonids
!
The new species was Pentanymphon antarcticum, whose genericname indicates that it differs from the well-established octopodousNymphon only in the possession of an additional pair of legs. Anotherspecies, P. minutum, has recently been described (Gordon, 1944)'Shortly after the discovery of Pentanymphon, a third type, relatedto the octopodous Pycnogonum, was found, one species in the Antarcticand another, strangely enough, from French Guiana, and a secondspecies of Decolopoda was collected in the Antarctic. As if this werenot enough, a 12-legged specimen, which has been named Dodeco-lopoda mawsoni, was found by the recent British, Australian, andNew Zealand Antarctic Research Expedition on the edge of theAntarctic south of Kerguelen. Finally, among the material belongingto the United States National Museum, I found several specimens ofstill another form of lo-legged pycnogonid, which I have namedPentacolossendeis reticulata, collected as early as 1872 off the FloridaKeys. It is something of a mystery why this species remained unde-scribed so long, since it was first collected by William Stimpson,the United States of America (Nature, vol. 146, pp. 238-239, 1940), and Con-gressional Record, vol. 86, Appendix, pp. 3I94-3I95, 1940.Now, after more than a hundred years, one of Eights' original specimens ofDecolopoda has been found among the collections in the Museum of ComparativeZoology. Although the only information available concerning this specimen isa cryptic "Parchment No. 952" and a catalog entry, "South Shetlands," thecircumstances indicate that this is one of the long-lost types. It is possible thatit was presented to Dr. Amos Binney, together with the manuscript and platereferred to in the letter from Eights to Dr. Binney (included in Caiman'spaper), and eventually found its way to the Museum of Comparative Zoology.The specimen is lacking a few of the terminal joints of some of the legs andone of the ovigers is detached, but it is otherwise in good condition. Accordingly,the specimen has been designated "Neoholotype, d*, M.C.Z. No. 1227 1."
NO. l8 PYCNOGONIDA—IIEDGPETII II
author of several species of pycnogonicls, and although he died beforefinding time to study his material, the species was collected again in1893 by the State University of Iowa expedition, and three morespecimens were collected by the Fish Hazvk in 1902.''There are, then, seven, or perhaps eight (there may be a thirdspecies of Decolopoda), species of decapodous pycnogonids, and onedodecapodous species. They are so far known only from the Ant-arctic and American tropical regions, and several of them are com-mon, to the extent that every expedition manages to collect severalspecimens. They are neither isolated freaks nor monstrosities, butrelatively stable forms.The most curious thing about these extra-legged pycnogonids istheir close resemblance to certain "normal" genera, a resemblancewhich in some cases extends to a particular species. Pentanyiiiphonis simply a Nymphon with an extra pair of legs, Pentapycnon a 10-legged Pycnogonum, and Pentacolossendeis would be Colossendciswithout its extra legs. Decolopoda, however, is somewhat differentfrom the thick-set species of Colossendeis which it resembles in thatit possesses chelifores, but since chelifores are occasionally retainedthrough the last m.oult stage in some individuals of the genus, thedifference is not as great as it seems. Dodecolopoda is merely anextra-legged Decolopoda, and is so far known only from a singlespecimen.The most conspicuous example of resemblance between a decapo-dous and an octopodous form is that of Pentapycnon charcoti Bouvierand Pycnogonum gaini Bouvier. Both of these are Antarctic forms:P. charcoti occurs in the South Shetlands, and P, gaini has been col-lected from the Palmer Archipelago, Ross Sea, and eastward to 54° E.Bouvier, who described both species, commented upon the similarities
* William Stimpson, M. D. (1832-1872), was "a naturalist of no mean capacity"who gathered a fine collection, wrote largely in Latin and was director of theChicago Academy of Sciences from 1865 to 1872. He lost all his work andcollections (including the Pourtales collection from the Florida Keys) in theChicago fire of 1871, and never recovered from the shock. In April 1872 hewent to the Keys on the Bache, but even this did not revive him and he diedon May 26. Nathaniel Southgate Shaler, in his Autobiography, pp. 128-129,has an interesting little story about Stimpson. It happened in those days whenAgassiz pere reigned at Harvard and "that Darwinian hypothesis" was not tobe mentioned except in private. According to Shaler, Stimpson "was muchpuzzled by the transitional varieties between many of the species of molluscs hewas studying, especially those occuring among the fresh water gastropods. Onone occasion I saw him throw one of these vexatious shapes upon the floor,after he had studied it for a long time, put his heel upon it and grind it to powder,remarking, 'That's the proper way to serve a damned transitional form.'
"
60'W
60' \V
Fig. 4.—Distribution maps of polymerous pycnogonids.12
NO. l8 PYCNOGONIDA—HEDGPETH I3between them at some length (1913, pp. 157-160), but placed stresson the last dorsal trunk tubercle of F. gaini as a vestigial remnant ofthe lost fourth somite, in support of his theory that the lo-leggedspecies were the primitive forms. The essential difiference betweenthese two forms, according to Bouvier's figures, is the dilated pro-boscis, adorned with a dorsal tubercle, of Pentapycnon charcoti. Ac-cording to Gordon (1944, p. 69), the proboscis of Pycnogonum gainiis sometimes dilated at the tip and may also bear a noticeable tubercle.It would seem, then, that these two forms are closely similar, withessentially the same range of variation, and that it would be impos-sible to refer a specimen lacking the posterior segments to its "genus."This same parallelism is evident, but not as pronounced, betweenthe tropical American Pentapycnon geayi Bouvier and a West Indianspecies, Pycnogonum sp.^ Here again the most conspicuous differencebetween the two forms is the shape of the proboscis, which is longerand more tapered in the decapodous form. Also, its do?sal tuberclesare taller. The ovigers are almost identical and could not be toldapart if separated from the specimens. Unfortunately there is notenough material available of either species to determine the range ofvariation.Turning to the lo-legged nymphons, we find similar examples ofpaired species. Although it might be protested that the genusNymphon is such a large and complex one that it would be easy tofind an approximate counterpart of a lo-legged form, the case isstrengthened by the existence of a double parallel, in which the twodecapodous forms are closely related to a pair of octopodous species,Nymphon hiemale Hodgson and Nymphon gracillimum Caiman.The most widely distributed decapodous form, Pentanymphon ant-arcticum,^ differs from its cognate "normal" species, Nymphon hie-
5 This is an unpublished species. I regret that this discussion necessitatesmention of this species before its formal description in my forthcoming mono-graph of Western Atlantic and Caribbean species.
^ If the future taxonomists act on the suggestion that such decapodous generaas Pentanymphon, Pentapycnon, and Pentacolossendeis cannot stand alone, itshould be noted that both antarcticiim (Miers, 1879), and minutum (Goodsir,1842) have been used for Nymphon and that nomina nova might be requiredfor the pentamerous forms, since these names would be unavailable for trinomialdesignations. As a matter of convenience, the pseudogeneric names should beretained, but in any event, names for new decapodous species should be differentfrom those in the respective octopodous genera. Dr. Hobart M. Smith (Science,vol. 102, No. 2643, pp. 185-189, 1945, Categories of species names in zoology)has proposed an elaborate classification of species names, which does not, how-ever, suggest a solution for this particular problem. It might be feasible, if it
14 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. I06
Fig. S.—a, Pycnogomim gaini; b, Pentapycnon charcoti; c, Pycnogonum sp.
;
d, Pentapycnon geayi. (a and b, and oviger of P. geayi after Bouvier, 1913; c andd, original and to same scale.)
male, in the following respects: The tarsus of F. mitarcticum issomewhat shorter than the same joint in A^. hiemale, the third jointof the palpus may be slightly longer, and the compound spines of theturns out these forms actually represent the same species, to adopt somesort of exponential notation, such as Nymphom hiemale [antarcticum], or N.hiemale^, etc. It does not seem to me that these pseudogeneric names can beused as subgeneric categories, as the forms represented do not conform withthe usual conception of a subgenus.
NO. I< PYCNOGONIDA HEDGPETH 15
oviger are somewhat different (see fig. 6). According to Hodgson'sfigures, the differences between these compound spines are conspicu-ous, but in the two specimens I have examined they are not so signifi-
1 ZFig. 6.
—
Nymphon hiemale and Pentanymphon antarcticum. a, dorsal view oftrunk ; b, terminal joints of oviger ; c, chela ; d, tarsus and propodus ; e, compoundspines of oviger; f, compound spines of oviger (after Hodgson, 1907): i, P.antarcticum, 2, N. hiemale. Scale of magnification for each pair of structures isthe same.
cant. Evidently there is some variation in the conformation of thesespines with individual specimens, for the differences are too great tobe explained by artistic interpretation of a minute structure whose
i6 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. I06location renders large-scale camera-lucida drawings impossible. Thetwo species are conspicuously alike in general appearance and con-formation of the chelae.The other decapodous species, P. miniitum, is a smaller edition ofP. antarcticum, and is very close to N. gracillimum, which in turnappears to be a smaller form of A^. hiemale. It is closer to N. gracil-limum than P. antarcticum is to A^. hiemale, especially in the structureof the compound spines and proportions of the tarsal joints. Sinceall four of these forms are highly variable in these and other details,Table 2.—Ratios and measurements of Nymphon and Pentanyphon(Compiled principally from Gordon, 1932.)
NO. l8 PYCNOGONIDA—HEDGPETH 17
specimen of Pentanymphon minutum places it outside this trend.However, if it had happened to be twice as large, its position on thegraph would be precisely where expected, namely, in the same rela-tion to Pentanymphon antarcticxim as NympJion gracillimtini is toA^. hiemale. Therefore it can be assumed that the trend of the leg-trunk ratio of Pentanymphon minutum is parallel to that of the otherthree forms.This close relationship between decapodous and octopodous formsis not so evident for the species of Dccolopoda, since the retentionof the cheli fores sets them apart from Colosscndeis at the outset.However, the Colossendeis most closely resembling Decolopoda, C.
-4—E
/? antarcticum(position ofP minutum iF5136 uyere doubled)' N. hiewale/\I.graciIIimu/r)o
o
.--(presumed trend d\ ratio)
'-^^/pTVinutum
log. length of leg
30°
Fig. 7.-
1
—
\
—
\
—
\ rr^20 mm 30 W 50 60 70 80 70
-Logarithmic graph of the ratios of NympJion and Pentanymphon.
tvilsoni, shows a close agreement in proportions with D. antarctica,and this resemblance is further emphasized by the fact that C. wilsonihas eight-jointed palpi, instead of the usual nine for the genus. Fromtable 3, giving ratios and comparison of anatomical characters, itcan be seen that C. wilsoni agrees more closely with D. antarctica thanwith D. australis except in the leg-trunk ratio, in which respect itagrees with the specimen of Decolopoda from Heard Island measuredby Gordon (1944). It is of interest to note that this Heard Islandspecimen, which is the nearest record for Decolopoda to the type andonly known locality for Dodecolopoda, shows more similarities toDodecolopoda than to these species of Decolopoda from the Americanquadrant of the Antarctic. The intermediate character of the ratios ofColossendeis zvilsoni, as contrasted with those of Decolopoda antarcticafrom the Antarctic archipelago and D. sp. from Heard Island, is of
SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. I06
Fig. 8.—a, Colossendeis wilsoni (after Caiman, 1915) ; b, Decolopoda australis:c, Dodecolopoda mawsoni (after Caiman and Gordon, 1932). Dorsal and lateralviews of trunk. All drawings to the same scale.
NO. 11 PYCNOGONIDA—HEDGPETII 19further interest in view of the occurrence of C. zmlsoni at Cape Adare,about midway between those locaHties.As Caiman and Gordon (1932, p. no) pointed out, the occurrenceof a pycnogonid with six trunk somites "does not really involve anyimportant modification of the problem presented by the ten-leggedspecies." As table 3 shows, the widest divergence between Dodeco-Table 3.
—
Anatomical characters and measurements, Colossendeis,Decolopoda and Dodecolopoda(Compiled principally from Gordon, 1932 and 1944.)
Colossendeiswilsoni Decolopodaantarctica Decolopodaaustralis Decolopodasp. (HeardIsland) DodecolopodamawsoniTrunk:length, mmarea\/ areawidth, 2d lat. proc.legPalpus, No. of jts.Chelifore:length, 1st jt.length of trunk"length 1st jt.width 1st jt.chela
Eye tubercle
.
Leg:length, mm.tibia Ifemurtibia 2femurlegtrunk
5.28283 03
•94
more thanhalf widthof cephalicseg.34-391 .06
I 256.5
.
96-1 . 008-10
.66-. 756 . 00-8 . GOpalm long,fingersslightlyarchedmore thanhalf widthof cephalicseg.
I .05-1 .08
I. 22-1. 33lO.-II
.
9-10*553-8
.85-. 929
•5-. 6540-5.
3
palm short,fingersstronglyarchedless thanhalf widthof cephalicseg.90-100*
I . oo-i . 04
I. 04-1. 13lO.-II
.
•979
.463^0palm short,fingersstronglyarchedless thanhalf widthof cephalicseg.
1 . 121.245 40
181485^28
•755-4palm short,fingersstronglyarchedless thanhalf widthof cephalicseg.240.31. 19
1 .2013-4
* Estimated from photograph of neoholotype.lopoda and Decolopoda is the leg-trunk ratio. Unlike the decapodousnymphons, in which the body length is materially increased by theaddition of a fifth somite, Decolopoda and Dodecolopoda appear toincur no noticeable increase in body length with the addition of so-mites over the ratio for the closely related C. wilsoni. Recognizingthat their species might just as easily be called a Decolopoda, Caimanproposed a new generic name for it simply as a taxonomic conveni-
20 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 106
ence, and it is certain that Dodecolopoda represents one more step inthe pattern of variation beyond Decolopoda. Its cognate decapodousform is probably Decolopoda australis. UnUke the decapodous formsof the other genera, these polymerous colossendeids are much largerthan the related octopodous form.This larger size, however, is not disproportionate, but appears torepresent an arithmetic progression between Colossendeis wilsoni,Decolopoda australis and Dodecolopoda mawsoni. Furthermore,when the ratio of trunk length to leg length is plotted against the
NO. l8 PYCNOGONIDA HEDGPETH 21tween C. wilsoni and D. mazvsoni. A further coincidence is the indica-tion that the curve for leg-trunk ratio of the Nymphon-Pentanymphongroup is also at an angle of 30°, although it has a downward ratherthan an upward trend.More precise and extensive data might reveal some interesting factsabout ratios and growth rates, especially if a growth series could beassembled. It would be particularly interesting to verify theNymphon-Pentanymphon curve, and confirm the apparent trend of
Dodecolopoda mat^jonl -&—
c Decolopoda australis-^^
•i^entapymphon avtarcticum
,.''1 ^:^=^ympnon qracillimum
-5 ,'''30° ^'•-.
^ ^^4^Pentanymphon minurum "'-.^
loq. lenqlb of leq30 mm 60 100 150 160 200 230Fig. 10.—Combined logarithmic graph of trunk-leg ratios of Nymphon, Penta-nyinphon, Colossciulcis, Decolopoda, and Dodecolopoda.30° for the polymerous forms in both families. It is not certain thatmuch can be proved by such analysis, other than to demonstrate acommon set of numerical values for the different types of polymerousforms. One should be wary of inferring too much from the trendsrevealed by logarithmic plotting, for they are inherent in the methoditself. It is easy to become bemused by these pretty graphs, and theyhave fascinated several biologists to the extent that their contributionson the subject might be termed a logarithmic analysis of the Adyo«.Although Pcntacolossendeis reticulata is not a rare species, andmay presumably be collected almost at will along the hundred-fathomline south of the Florida Keys, no closely related Colossendeis from
22 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. I06
the same region has come to light. It is of particular importance,however, in demonstrating that the occurrence of lo-legged pycnog-onids in the tropical American region is not an isolated phenomenon,confined to a single species. It is also of further interest in that intwo of the five known specimens the second trunk somites are smaller,and the second pair of legs arising from this somite are slightlyshorter than the remaining legs. Since this difference occurs not once,but twice, it would seem that this is not an individual variation, butis in some way correlated with the decapodous condition.It is difficult to ascribe any external environmental cause to theorigin of the decapodous (and dodecapodous) forms, for two morediverse sets of conditions than those prevailing in the Antarctic andthe American subtropical or Caribbean regions could not be imagined.These regions are widely divergent in both salinity and temperature.For example, the salinity of the Antarctic region of the South Atlanticranges from 33 to 35 parts per thousand, whereas that of the tropicalAmerican region has a range of 36 to 38 parts per thousand. Indeed,the only physical condition these areas appear to have in common isthat both overlie tectonic arcs, areas of stress in the earth's crust wherenegative anomalies in gravity may occur.^ However, this geophysicalcondition shows little correlation with the occurrence of decapodouspycnogonids, since they are unknown from the East Indies, Japan, andother island arc regions. If future collections do bring more decapo-dous pycnogonids to light, however, it is safe to suggest that the mostlikely area in which they will be found is the East Indies.It may be significant, for reasons still not apparent, that decapodouspycnogonids are most numerous, both in species and numbers, alongthe Antarctic arc from the Palmer Peninsula to South Georgia.Furthermore, within this critical area the decapodous forms andtheir corresponding octopodous forms occur only south of the zone ofAntarctic-South Atlantic convergence. Although the available dataare not extensive, there seems to be a further limitation of decapo-dous forms to the colder waters of the Antarctic, whereas the cor-responding octopodous forms, at least in the genus Nymphon, occurin higher temperatures. This becomes evident when the distributionof the Nymphon-comp\e:x. is mapped against the temperature distri-bution (fig. 11). Of course it is also true that a similar correlationcan be assumed for salinity, but in this case the range does not seem
s Measurements of isostasy have not yet been made in the Antarctic, but Hess(Proc. Amer. Philos. Soc, vol. 79, No. i, p. 73, 1938) suggests that a negativeanomaly strip "will almost be certainly present around the Cape Horn, SouthGeorgia-Antarctica island arc."
NO. I( PYCNOGONIDA HEDGPETH
to be great enough to justify any generalization, inasnuicli as it is inthe magnitude of 00.3 parts per thousand.Because of the virtually identical distribution of the two species ofDecolopoda, no generalizations as to their distribution can be inferred.As for Pentapycnon charcoti and Pycnogonum gaini, it would seem,in this case, that the decapodous form is the more northern one, forit was found in the South Shetlands, north of the Palmer Peninsula,whereas Pycnogonum gaini occurs near the base of the peninsula, inthe Ross Sea, and along the edge of the Antarctica south of NewZealand and Australia. This distribution pattern is tentative, inas-
A Nymplioiiliiemale. ^ Pentanympbon antarcticurn.Fig. II.—Distribution of Nymphon-Pentanymphon compared with temperatureof 100-meter surface layer. (Isotherms from Deacon, A general account of thehydrology of the South Atlantic Ocean, Discovery Rep., vol. 7, fig. 12, 1933.)much as Pentapycnon charcoti is so far known from a singlecollection.The distribution of the warm-water forms in the American sub-tropical region is inadequately known, despite the greater accessibilityof the area to collectors, and a comprehensive hydrography of theregion is yet to be worked out. Hence little can be said about distri-bution in this area that cannot be inferred from an inspection of themap (fig. 4, A) which indicates all the known localities for decapo-dous forms, as well as for the closely related Pycnogonum sp.Once the existence of lo-legged forms had been established, theinevitable discussion as to their phylogenetic significance got under
24 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. I06
way. Those who participated in the argument immediately dividedthemselves into opposing camps: the proponents of the theory thatthe lo-legged condition represented the original, primitive state ofthe Pycnogonida, and those who believed it to be a secondary phenom-enon arising out of the octopodous condition. The first to suggestthe primitive nature of the decapodous condition was Cole (1905),who argued that Decolopoda represented the progenitor of two di-verging series of phylogenetic lines, leading to Pycnogonum on onehand and to Colossendeis on the other. The most persistent advocateof this position was Bouvier, who maintained his belief in the primi-tive origin of the decapodous type in his last paper on Pycnogonida(1937), in spite of the discovery of Dodecolopoda, for he consideredthe five ventral ganglia of the octopodous pyncogonid an indication ofthe original number of trunk somites. If that were the case, we mightexpect a radical change in the anterior ganglia of decapodous forms,but such is not the fact. In Decolopoda there is simply one moreventral ganglion added to the end of the chain, and the enervation ofthe cephalic region remains unchanged. (See Gordon, 1932, pp. 128-130, %• 7Z-)Bouvier placed particular emphasis upon a larva described byDogiel (1911) from the Murman Station in the Arctic as evidence infavor of the primitive character of the decapodous condition. Thislarva (see fig. 12, c) of Nymphon spinoswn was fairly well advanced,and possessed a fifth pair of rudimentary legs on the posterior seg-ment. Dogiel believed this to be an atavistic deformity, but it seemsmore likely that it was simply an isolated example of faulty develop-ment. If it were actually a throw-back, we should expect it to have anindication of the fifth segment, which it does not have, in the illus-tration at least, and we might also expect it to be a more commonoccurrence. Dogiel's example is the only one recorded in the litera-ture. For that matter, anomalies and deformities seem to be rareamong the Pycnogonida, aside from those caused by the regenerationof lost parts. The most conspicuous one I have encountered is aspecimen of Achelia borealis from the North Pacific, which has butthree legs on the right side. (Fig. 13, b.) There is no evidence oftraumatic injury in this specimen and it appears to be a congenitaldeformity. What is apparently the result of regeneration is describedby Schimkewitsch and Dogiel (1913) in a specimen of Anoplodactyluspetiolatus from Millport, Scotland (fig. 13, c). Bouvier (1914) ex-amined a collection of 3,268 specimens of Pycnogonum littorale fromPlymouth, England, and found only one abnormal specimen, a femalewith seven legs, the last pair being replaced by a median one. Bouvier
NO. l8 PYCNOGONIDA—HEDGPETH 25
considered this deformity to be the result of an injury at a fairlyearly stage.Two indubitable examples of congenital abnormalities have recently
d
Fig. 12.—a, Protonymphon larva of Achelia echinata (after Dohrn, 1881) ;b, protonymphon larva of Pentanymphon antarcticum ; c, larva of Nymphonspinosum (after Dogiel, 1911) ; d, larva of Nymphonella tapetis (after Ohshima,1942b).been described by Ohshima (1942a, b). The first of these is a speci-men of Callipallene brevirostris from Sasebo, in which there are butthree pairs of walking legs. Otherwise the specimen is a perfectlyformed male, bearing eggs. There are four pairs of trunk ganglia.
26 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. I06
Fig. 13.—Ventral view of trunk of six-legged specimen of Callipallene breviros-tris (after Ohshima, 1942a) ; b, ventral view of trunk of seven-legged specimenof Achelia borealis (from Albatross station 5037, 1906) ; c, six-legged Anoplodac-tylus petiolatus with trifid last leg (after Heifer and Schlottke, fig. 142) ; d, six-legged Nymphonella tapctis with trifid leg (after Ohshima, 1942b).
NO. l8 PYCNOGONIDA HEDGPETH 27
the fourth also serving- the third pair of legs so that this pair receivesa double set of nerves (fig. 13, a). Obviously the fourth pair of legsis missing here. It is interesting to note that the angle of the thirdpair of lateral processes in this specimen is approximately the meanvalue of the angles of the third and fourth processes of a normal indi-vidual. The second abnormality described by Ohshima is more com-plicated. In this case it is a six-legged advanced larva of Nymphonellatapetis, a form in which the legs arise by simultaneous budding ratherthan by addition from the anterior region in successive molts, as inmost other pycnogonids in which the larval stages have been observed.On the left side, the second leg receives nerves from the second andthird trunk ganglia, whereas on the right side the third and fourthtrunk ganglia serve the third leg, which is tri furcate distally (fig.13, d). This abnormal distal branching of appendages is not rareamong arthropods, and there are several reported examples of itsoccurrence in pycnogonids independent of abnormalities in segmen-tation. (See Gordon, 1932, pp. 130- 131.) It appears to have norelation to the problem of polymerism, although the duplication ofnerve supply may have induced it in the examples described bySchimkewitsch and Dogiel, and Ohshima.In the seven-legged specimen of Achelia borealis, the odd leg onthe right side receives the nerves from the ganglia which serve thesecond and third legs on the other side (fig. 13, b). This median legis so located that it balances the second and third legs of the normalside. Like Ohshima's six-legged specimen of Callipallene hrevirostris,this anomalous specimen is an ovigerous male. Ohshima (1942b) sug-gested that the aberrant specimen of Nymphonella tapetis may havebeen formed by the failure of the limb buds to divide, but as heanticipated, this would not account for a similar abnormality in aform in which the legs were not formed in this manner. Probablythe difference in larval development between Nymphonella and otherpycnogonids is actually not as great as he seems to believe.From these anomalies it is apparent that the occurrence of fourtrunk ganglia is very stable. They would also seem to indicate thatthe loss of ganglia is a rare occurrence, in contradiction to Bouvier'ssuggestion that the octopodous forms have lost a trunk ganglion. Itseems more likely that it is easier for a pycnogonid to add ganglia thanto discard them, although several forms have post-trunk ganglia inthe larvae which are coalesced with the last trunk ganglia in the adultforms, and there is a tendency toward fusion of the anterior trunkganglia and cephalic ganglia in the compact, disciform types.As early as 1905, G. H. Carpenter expressed the opinion that the
28 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. Io6decapodous condition was a secondary modification, and he waspromptly seconded by Caiman in 1909, who reaffirmed his position inhis Terra Nova report (1915) as a rebuttal of Bouvier's contentionsin behalf of the primitive character of the decapodous condition.Caiman also denied Bouvier's suggestion that the phenomenon couldbe localized ; it was Bouvier's belief that the octopodous forms hadlost the fourth trunk somite. Although Caiman did not believe thatthe extra segment could be so precisely localized, his suggestion (Cai-man and Gordon, 1932, p. iii) that "the metameric instability whichwe believe to have affected the trunk somites may possibly have in-fluenced the segmentation of [the palpus of Decolopoda antarctka],"contains a hint that the instability may be primarily effective in theanterior region. This possibility cannot be ignored, although Dr. Cai-man has assured me, in litteris, that he does not believe it can be solocalized. Moreover, the fact remains that the major differencesamong the families and genera involve the varying combinations ofanterior appendages. This is further suggested by Nymphonella, inwhich the essential difference between it and the closely related Asco-rhynchus is the secondary segmentation of the palpus and first pair oflegs. Also, the smaller second trunk somites of Pentacolossendeisreticulata would seem to indicate that the extra somite, in this decapo-dous form at least, arose in the anterior region. On the other hand,instability in the last trunk segment is suggested by Ohshima's six-legged specimen, and he (1942a, p. 260) is of the opinion that thepentamerous forms arose through such instabihty: "Thus either theincrease or decrease in the number of body segments, and conse-quently in appendages, takes place at the junction of the trunk andthe tail (abdomen), but not as hypertrophy or abortion occurring atthe morphological posterior end of body."The possible localization of this phenomenon in a particular regionis not the major problem, however. Even if that could be satisfactorilyanswered, the question still remains : what, exactly, is the nature, thecause, and significance of the decapodous condition in the Pycno-gonida? It is a phenomenon without counterpart in any other knowngroup of animals, and the various attempts to compare it with thesupernumerary pregenital somites of Polyartemia (Caiman, 1915)and the additional gill arches of the shark Pliotrema are of little morethan academic interest.The uniqueness of the phenomenon can be appreciated when it isremembered that it occurs in three widely divergent family types, yetis at the same time closely correlated with particular species or speciescomplexes. Furthermore, it is apparently correlated with the evolu-
NO. li PYCNOGONIDA—HEDGPETH 29
tionary force which governs variation within the group as a whole,since it occurs in those groups in which variation is confined to thespecific rather^ than to the generic rank. As a corollary to this, it isinteresting to note that the only established example of supernumerarysegmentation in other branches occurs in the Ammotheidae (in which,incidentally, the number of joints of the palpus varies, often within thegenus), namely, the reduplicated segmentation of the palpus and firstpair of legs of Nymphonella.^It should also be pointed out that this phenomenon of reduplicatedsegmentation, or polymerism, occurs in those branches of the Pycno-gonida which may be considered, because of their large numbers ofnarrowly separated and numerically abundant species, as the mostsuccessful from the evolutionary standpoint. In other words, the
Fig. 14.
—
Ascorhynchiis ramipes (after Lou, 1936) ; b, Nymphonella tapefis(after Ohshima, 1935b).
maintenance of generic form throughout a large series of species incertain branches or families is a possible symptom of dynamic tension
—Suva/At?, as Aristotle would have called it—and when the tension ishigh, extra-legged forms are the result.^ ° Conversely, in those groupswhere the wide divergence of generic pattern and a correspondinglylow ratio of species to genus may indicate a low dynamic potential,the basic metameric stability is not upset. It is of further interest tonote that, among the Antarctic species at least, the lo-legged forms
9 Bouvier (1910) reported a specimen of Ettrycyde with 17 joints in the palpus,but this appears to have been an individual anomaly. This genus is closely re-lated to Ascorhynchiis.10 For a discussion of the dynamics of evolution, see Lotka, Alfred J., Elementsof physical biology (Williams & Wilkins, Baltimore, 1925), chaps. 2-4. Lotkadefines evolution as "the history of a system undergoing irreversible changes."(P. 24.)
30 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. I06
are more abundant and more widely distributed than the correspond-ing 8-legged forms. Incidentally, the relatively greater success of thedecapodous forms should be another point against the theory ofprimitive origin, for evolution does not go backward (although itmay sometimes stand still) and there would be no conceivable ad-vantage in reverting to a primitive type once the octopodous typehad proved so successful.The characters of the decapodous pycnogonids, their close resem-blance to particular species, greater success as organisms (as indicatedby their wider distribution and abundance vis-a-vis the cognate octo-podous forms), and overlapping but not precisely identical distribu-tion, suggest that they are polymorphic forms of the octopodousspecies. ^^ This cannot be proved until studies of the chromosomes areavailable, but it seems the most plausible explanation in this day andage when chromosomes are quite the fashion. Certainly it is temptingto suggest that decapodous forms are the immediate result of doubledchromosomes and that the dodecapodous form is a possible tetraploidtype. There is some support for the suggestion that this is a poly-ploid condition in the fact that the lo-legged forms occur in whatare probably the maximum and minimum temperature ranges forpycnogonids. Temperature extremes appear to induce polyploidy, par-ticularly in plants (cf. Huxley, op. cit., p. 337).Unfortunately, live material of the species involved is inaccessible tolaboratory workers, and, for that matter, the normal chromosomenumber of any pycnogonid is yet to be determined. Furthermore,polymorphism is not necessarily a result of polyploidy, complete orpartial, and can only be finally determined by the discovery of both8- and lo-legged forms in a single brood or from successive or alter-nating broods of a single female. Hence, before this problem canbe adequately investigated, it will be necessary to determine the chro-mosome number of common species as well as of those involved in thelo-legged problem, and to develop laboratory culture of livingmaterial.^^
1^ For an extended discussion of polymorphism, see Ford, E.B., Polymorphismand taxonomy, in The New Systematics, pp. 493-513, and Huxley, Julian,Evolution: The modern synthesis, especially p. 96 et. seq.
^- Of course, as Goldschmidt maintains, it is possible that chromosome differ-ences may or may not indicate anything, and that the chromosome patternmay change without visible effect on the genotype, but Goldschmidt's heresiesare not well received in the strongholds of the chromosome cartographers. (SeeGoldschmidt, Richard, The material basis of evolution, Yale University Press,1940, especially pp. 186 and 191.) As for laboratory culture, it is probable thatit will prove to be relatively easy. Dohrn (1881) kept an amputated specimen
NO. l8 PYCNOGONIDA HEDGPETH 3ISuch laboratory investigations may demonstrate that metameric re-duplication among the Pycnogonida is a completely different type ofvariation than heretofore known, but whatever its mechanism, thefact remains that it is too intimately bound up with particular speciesto be a random coincidence or genetic accident. It may be discoveredthat the basic chromosome pattern of the three families in which itoccurs is identical and possibly different from that of the otherfamilies. The success of this variation, as indicated by its relativeabundance, indicates that it is in some way advantageous, although wemay not be able to perceive wherein the advantage lies. The lateDr. C. Tate Regan, participating in a discussion of this problem at ameeting of the Linnean Society, remarked that Dodecolopoda and thedecapodous forms appeared to be an example of "evolution by acci-dent, a phenomenon difficult to understand." ^^ Possibly he had inmind the same difficulty which led Aristotle to deny that variationcould be accidental (and hence infinite) : "Nature, however, avoidswhat is infinite, because the infinite lacks completion and finality,whereas that is what nature always seeks." ^*Undoubtedly William Morton Wheeler would have considered thepolymerous Pycnogonida an example of emergence, which he wascareful to restrict to its "epigenetic" meaning, as distinct from theall-inclusive sense (with its overtones of spiritual emergence, crea-tive evolution, elan vital and the rest of it) of less realistic biologistsand philosophers. Emergence would indeed be a handy name for thisphenomenon, but it is little more than a name, and with all deferenceto the late Dr. Wheeler, a rather dangerous name, because of itsphilosophical aura. As a concept, emergence now has little sanction,either in biology or philosophy.^^
alive for 4 weeks while observing regeneration, and Arita (1937) kept a colonyalive in flowing sea-water for 10 weeks. Specimens of pycnogonids collectedon the shore often live in small jars for a day or more without a change ofwater. Temperature control, especially for cold water species, may be as im-portant as salinity and oxygen conditions.12 Cf. Proc. Linn. Soc. London, Sess. 145, 1932-33, pt. 2, pp. 91-93.1* Gen. Anim. I, i. (715 b, 15-16), Loeb Classics ed., pp. 6-7:7) de (fiixTis <pevyei to aireipoy to /lev yap aneipov dreXes, i] de <f>vais del f^jret reXoj.
^^ For a statement of Wheeler's position, see Emergent evolution and thedevelopment of societies, in Essays in Philosophical Biology, pp. 143-169 (Har-vard Univ. Press, 1939). Julian Huxley, in his Evolution: The modern syn-thesis, does not even mention emergence, and denies the need for postulating an
"elan vital" (p. 568). For the present philosophical status of the concept, seeIrwin Edman's introduction to the Modern Library edition of Bergson's Creativeevolution.
32 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. I06
Nevertheless, this inferentially invites the ghosts of teleology intothe discussion, but they implicitly haunt all speculations in theoreticalbiology, and no one, including the late Dr. Wheeler, despite his unkindremarks about the neo-Thomists, has yet discovered an efficaciousformula of exorcism. The whole range of form and variation withinthe Pycnogonida is a compact, integrated pattern, and patterns arenot aimless accidents induced by genes behaving like Mexican jump-ing beans on a warm day. No one has done more to show that growthand form are achieved in conformity with physical laws than thatenthusiastic student of Aristotle, D'Arcy Wentworth Thompson, andif any inference can be drawn from his classic monograph, "OnGrowth and Form," it is that teleology is far from being a dead con-cept and that at least one purpose of an organism or group of organ-isms is adaptation to and exploitation of its environment to the limitof its capacity to utilize physical laws.^® True, this can be construedsimply as a description of the evolutionary process without invokingthe Aristotelean tcAo?, but so clear a process as evolutionary adapta-tion implies a Cause. Dr. Julian Huxley suggests that the purpose, orCause, of evolution is Progress, and perhaps this is as good a guessas any, although some will protest that it still leaves us within thephilosophical circle without a clear way out.^'^16 On growth and form (Macmillan Co., New York, new ed., 1942). "Still,all the while, like warp and woof, mechanism and teleology are interwoventogether, and we must not cleave to the one nor despise the other, for their unionis rooted, in the very nature of totality." (P. 7.) See also Lotka, op. cit., chap. 9.1^ Huxley, op. cit., chap. 10. It seems to me that "progress" is an unfortunateterm for Dr. Huxley's conception of evolutionary development. Furthermore,while remaining a staunch mechanist up to his last chapter, he inevitably com-mits that anthropomorphic and logical error of the mechanists, i.e., granted thatman is the inevitable result of evolutionary progress, he has now attained thepower to interfere with that mechanistic process of evolution which producedhim and direct its course so as to alter his own evolutionary future. "The futureof man, if it is to be progress and not merely a standstill or a degeneration, mustbe guided by a deliberate purpose." (P. 577.) This is tantamount to endowingman with the attributes of divinity, of being a First Cause within himself, andwhile this is not to deny that man is without the power to improve his racialstock by selective breeding, the full implications of this notion would tempt evena liberal clergyman in a university town to resort to St. Thomas Aquinas : "Itis possible for an effect to happen outside the order of some particular cause,but not outside the order of the universal cause." (Summa Theol. I, Q. 103, Art.7.) Of course, man cannot expect too much from St. Thomas, who deniessuch power to the angels (ibid., Q. 52, Art. 2), and says also: "For an indi-vidual man cannot be the cause of human nature absolutely, because he wouldthen be the cause of himself; but he is the cause that human nature exists inthe man begotten." (Ibid., Q. 45, Art. 5.) As an exercise in logic, it would be
NO. l8 PYCNOGONIDA—HEDGPETH 33But biologists are not alone in this philosophical dilemma: physi-cists, having pursued matter down to apparently anarchistic particlesthey call quanta, now find themselves again obliged to become philoso-phers and speculate upon First Causes, after what had seemed for atime a blessed emancipation from philosophy.^® As long as we searchfor an explanation for the nature of things as we find them in the natu-ral world, so long will we be haunted by teleology, and that will doubt-less be as long as man is on earth. Of course, it is dangerous to argueby analogy from the human mind, but the basic urge of all great intel-lects, be they scientists or philosophers, theologians or poets, toachieve unity out of the multiplicity of things known and perceived,suggests that Nature is up to the same thing in her endless adaptationsof diverse yet basically similar forms to the exigencies of the externalenvironment.^^
III. CONCERNING DISTRIBUTION AND DISPERSALAlthough the observation of Marcus (1940a, p. 197) that "theactive and passive means of distribution of the Pycnogonida seem tobe less than those in all other marine arthropods" is essentially true forlittoral species, there are several noteworthy examples of widespreaddistribution which are difficult to explain, and future collections,especially of the smaller forms, may prove many apparently localspecies to be widely distributed. This, however, would not vitiateinteresting to know on what grounds Dr. Huxley assumes that the future trendof human evolution is going to be static or even downhill, since he has assumedthat it has been "progressive" up to now. One might also infer that Dr. Huxley,like the late Dr. Wheeler, is more of a Lamarckian than he cares to admit inpublic, although the Chevalier's august name is mentioned in his book. Asfor progress, I will have more to say in Bios, March 1947, under the title"The Philosophic Jellyfish."18 Cf. Jeans, Sir James, Physics and philosophy (Macmillan Co., New York,1943). D'arcy Thompson is, naturally, fully aware of this difficulty: "More-over, the naturalist and the physicist will continue to speak of 'causes,' just asof old, though it may be with some mental reservations ; for, as a French philos-opher said in a kindred difficulty : 'ce sent la des manieres de s'exprimer, et sielles sont interdites il faut renoncer a parler de ces choses.' " (Op. cit., p. 9.)1^ Perhaps I do not sidestep "the vitalists, teleologists, et hoc genus omne"as adroitly as my former professor Dr. S. J. Holmes does in his paper, Theproblem of organic form (Sci. Montli., vol. 59, pp. 226-232, 253-260, 379-383.1944). Dr. Holmes discusses form as a result of chemical and physical equilibriaand interactions within the organism. His suggestion that life is "a ceaselessstriving for a peaceful heterogeneous equilibrium, the attainment of which wouldresult only in death" is not much different, philosophically, from my own state-ment, although I am limited by my material to external morphology.
34 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL, Io6Marcus' suggestion (idem) that the comparatively slight mobility ofthe Pycnogonida "may in some cases have favoured the developmentof the great number of minutely separated species," for the majorityof littoral species are probably limited to comparatively small areas.The most puzzling distribution is that of Ammothella bi-imgui-ciilata, originally described from the Bay of Naples. This littoralspecies is easily identifiable by the absence of the terminal claws ofthe legs, and all workers who have identified it from widely scatteredparts of the world have been obliging enough to supply adequatefigures, hence there can be little doubt that all the records, as widelyscattered as they are, represent the same species. Although most of
• Ammothella A fi.chelia echinata W Achclia spincsaFig. 15.—Distribution of Ammothella bi-nngiiiculata, Achelia echinata andA. spinosa. (Based on Goode Base Map No. 205, by permission of the Universityof Chicago Press.)these records have been described as geographical varieties, there donot seem to be enough anatomical differences among the various speci-mens to merit subspecific names. This species has been found alongthe shore of southern California, in Japan, Hawaii, and at RottnestIsland near Perth, Western Australia, in addition to the Bay ofNaples.Another widely distributed littoral species is Achelia echinata,which has been identified from northern Europe, the Bay of Naples,the Atlantic coast of Morocco, San Francisco, southern Alaska andthe Aleutians, Japan, the Siberian coast near Vladivostok, and Kiao-chow, China. There is also a closely related species from northeasternAmerica, Achelia spinosa, which some taxonomists have considered asynonym of ^. echinata, although it seems to me to be distinct enoughto merit specific rank. Nevertheless, it is probably a member of
NO. l8 PYCNOGONIDA HEDGPETH 35
the same species complex. The range of variation in A. cchinatais apparently large enough to justify a niunbcr of geographic varieties,and such a range of variation suggests that it is an older species thanAiiiiiiotJiclla hi-ungiiiculata. The distribution of Achclia echinata isthat of a typical Boreal species, and may represent a dispersal fromhigher latitudes as a result of the ice age.On the other hand, Anvmothella hi-imguiciilata is a warm-waterform whose distribution cannot be explained on such geologicalgrounds. Furthermore, the uniform character of the specimens fromvarious localities suggests that it is a young species. Unfortunately,we cannot tell whether or not this distribution antedated the sailingship with its bottom growth of hydroids and crannies in the hull inwhich such slow-moving organisms as pycnogonids might find refuge,but its pattern of dispersal suggests that sailing vessels had little to dowith the distribution of this species.-"Sporadic distribution, such as that of Aminothella hi-imguiculata, isnot a rare occurrence among marine invertebrates, including thoseforms with limited locomotive powers. The most conspicuous ex-ample to come to recent notice is that of a nemertean. Gorgorno-rhynchus, which is represented by closely allied species recently dis-covered in Bermuda and New South Wales. J. F. G. Wheeler, in anextended paper on this form, which dififers from all other nemerteansin the possession of a branched proboscis, advanced the suggestionthat the Australian and Bermudian forms arose simultaneously withinthe last few years, possibly by mutation, and that here was an exampleof evolution caught in the act. This rather extreme hypothesis over-looks,, as Zimmerman pointed out, the accidents of distribution andcollecting, and the possibility of fluctuating populations (at a lowcycle of abundance in the past it might easily have been overlooked).-^
20 Concerning the possibility of dispersal on vessel bottoms, this comment (inlitteris) by Dr. J. E. Benedict, Government Naturalist for the Falkland Islands,is interesting : "I have taken Caprella in a tow net in, roughly speaking, themiddle of the Atlantic. They were dead and could only have come from thefine bottom growth the ship had acquired in harbour in England." Shipwormsare often dispersed on wooden vessel bottoms. See, for example, Edmondson,C. H., Dispersal of shipworms among central Pacific islands, with descriptionsof new species, Occ. Pap. Bishop Mus., vol. i8. No. 15, pp. 211-224, 1946.
-1 Wheeler, J. F. G., The discovery of the nemertean Gorgornorhynchus andits bearing on evolutionary theory (Amer. Nat., vol. 76, pp. 470-493), and Zim-merman, E. C., On Wheeler's paper concerning evolution and the nemerteanGorgornorhynchus (ibid., vol. y7, pp. 27Z-2)7(>)- Coe, the nemertean authority,considers Wheeler's idea a "naive assumption." Cf. Coe, Wesley R., Thenemertean Gorgornorhynchus and the fluctuation of populations (ibid., vol.78, pp. 94-96).
36 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. I06However, the character of this variation, namely the longitudinaldivision of the proboscis, suggests that it may be analogous to thedecapodous condition in pycnogonids, itself a variation whose distribu-tion is curiously dispersed. When Wheeler suggests that Gorgorno-rhynchus is a simultaneous mutation in two widely separated parts ofthe world, caused perhaps by "an internal inevitable disruption ofsome sort," he takes two long steps ahead of his data and a hesitantsidestep toward Lamarckianism. However, it is certain that thereare conditions prevailing in some areas (such as the American Tropicsand the Antarctic insofar as pycnogonids are concerned) which in-duce speciation, or at least give a kind reception to variations. Itfollows that there must be a tendency within the organism to enableit to respond to those external conditions. If this be Lamarckianism,so be it. One suspects that many biologists have been browbeaten outof their sympathies toward Lamarckianism, for what is paleontologybut a long record of organisms which were capable or incapable ofresponding to changes in their environment, through the inheritance ofacquired or induced adaptations?Whatever the explanation for the distribution of Ammothella hi-unguiculata may be (future collections may prove it to be a circum-tropical species), the distribution of many small species in theNorth and South Atlantic can be explained on the assumption thatthe Sargassum provides a medium for their dispersal. At least ninesmall species are found on both sides of the Atlantic, and on theAmerican side six of these are found at Tortugas in the FloridaKeys. On the European side of the ocean these species are scatteredfrom Norway to Cape Verde, and the general pattern of distribu-tion suggests a dispersal from the American side of the Atlantic.At least two of these species, Anoplodactylns petiolatus and Endeisspinosa, are permanent members of the sargassum fauna in mid-Atlantic, and I have found Tanystylmn orbiculare, a species knownfrom Brazil and the United States, on Sargassum along the coastof Texas.^- The suggestion that the West Indian region may be acenter of dispersal for these various species gains some supportfrom the occurrence of identical and similar species on both sidesof the Isthmus of Panama. Perhaps more significant than the
2- Another method of dispersal is suggested by Lebour's (1916) discovery oflarvae of Anoplodactyhis petiolatus in the medusa stage of hydroids, at Ply-mouth, England. An excellent summary of the sargassum fauna will be found inthe paper by G. Timmermann, Biogeographische Untersuchungen fiber dieLebensgemeinschaft des treibenden Golfkrautes, Zeitschr. Morphol. and Oekol.Tiere, vol. 25, pp. 288-335, 1932.
NO. l8 PYCNOGONIDA—HEDGPETH 37identical species, from the standpoint of the distribution pattern, arethe pairs of closely related species, for three of the five Atlanticspecies with closely related Pacific species occur on both sides of theAtlantic. This may indicate that their occurrence in the westernAtlantic antedates their distribution to the eastern shores of theocean. This relationship is best illustrated in tabular form
:
Caribbean region Panamic region (chiefly)Callipalleneeinaciata* (Tortugas) californiensis (Southern California)Amjiwthellurugulosa (Brazil, Bermuda, Tortugas) heterosetosa (Galapagos)Ascorhynchusannatus* (Hatteras to Cuba) agassizi^ (Gulf of California)Eurycyderapliioslcr* (Tortugas) longisetosa (Colombia)Tanystylumorbiculare (Brazil to Massachusetts) duospinmn (Central California)
• On both sides of Atlantic.t Possibly synonymous with armatus.A curious aspect of the distribution of pycnogonids in the Atlanticis the occurrence of several species in Brazilian and European watersand their absence from northeast America and the West Indies.It is possible that this may be more apparent than real, for collec-tions from the northern shore of South America and the West Indiesare very inadequate, and several species described from Brazil haveturned up in collections from the West Indies.-^According to Ekman, the littoral fauna of the North Pacific issix to eight times as rich as that of the North Atlantic.-"* While itdoes not seem that this is altogether true for pycnogonids, this ele-ment of the fauna is more diversified in the North Pacific thanit is in the North Atlantic. There is but one endemic genus in theNorth Atlantic, Paranymphon, and that is a deep-water, not a littoral
-^ See Hedgpeth, 1943b. A more exhaustive discussion of western Atlanticand Caribbean species is now in press.
-* Ekman, Sven, Tiergeographie des Meeres, p. 231, Leipzig, 1935. An Englishversion of this work is now in preparation, under the direction of Karl P.Schmidt, of the Chicago Natural History Museum.
38 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. I06fomi. Another genus, Pigrogromitiis, is known by a single species
;
certainly it cannot be called an Atlantic form. There are at leastthree genera, Lccytliorhynchus, Nyrnphonella, and Dccachela, so farknown only from the North Pacific. The most striking differencebetween the pycnogonid faunas of the two oceans is the relativelyfew species widely distributed along both shores of the Pacific, incontradistinction to the considerable number of littoral species (in-cluding Boreal-Arctic forms), found on both sides of the Atlantic.As for the Boreal-Arctic species, it should be mentioned that rela-tively few of them are found in the North Pacific: Nymphon gros-sipes, N. longitarse, and Phoxichilidkmi femoratum are those most
Ti -Decachela \^ Lecythorhynchus H Nymphonella P Paranymphon ^ PiqrogromUus
l/linigma OAustrodecus Q) AustropaUene DAustroraplas 9Boehmia ^Discoarachne"^HannoniaVHeteronymphony.Oorhynchus 9 PycnothsaFig. 16.—Distribution of endemic genera of the Northern and SouthernHemispheres. (Because of dubious Northern Hemisphere records, Ammothea,the most abundant and widely distributed endemic genus of the Antarctic, hasbeen omitted.) (Based on Goode Base Map No. 205, by permission of the Uni-versity of Chicago Press.)
certainly present. Oddly enough, species of two of the endemicgenera are found on both sides of the Pacific : Lecythorhynchusmarginatus and Decachela discafa. Pycnosoma may also be endemicto the North Pacific, for Heifer's (1938) reference of a speciesfrom Chile to this genus is open to question (cf. Marcus, 1940b,p. 48). The California coast is especially rich in small littoral forms,there being perhaps 30 species in all identified from the coast be-tween Marin County (north of San Francisco) and San Diego.Although the littoral fauna of Japan is still incompletely knownit is evidently a rich one.-^ In addition to a large number of endemic25 In the collections made by the Albatross in 1900 and 1906, there are 18undescribed species, 6 of which were taken in shallow water. A systematicreport on these collections is now awaiting publication.
NO. l8 PYCNOGONIDA—HEDGPETH 39forms, it includes a strong element from the East Indies, which isnot found north of 35° N. lat., and a somewhat weaker representa-tion of the Pacific Boreal fauna north of 35°. The facies of thecombined littoral and shallow-water (less than 100 fathoms) faunais markedly different from that of the eastern part of the Pacific.One of the most conspicuous differences is the absence of Tany-stylmii from Japan and the northwestern Pacific as a whole, althoughthere are several species along the California coast. This divergencebetween the fauna of the western and eastern shores of the Pacificcan be explained in part by the lack of a convenient bridge offloating sargassum such as exists in the North Pacific. The intru-sion of large masses of Arctic water south of the Aleutian chain isprobably also an inhibiting factor, and in this connection it is inter-esting to note that the species found both in Japan and Californiaappear to be cold-water forms, with the exception of Amnwthellahi-iiuguiciilata.-^With the exception of the Antarctic, South Africa, and parts ofSouth America, the pycnogonids of the Southern Hemisphere areknown only from sporadic records, and much collecting remains tobe done before generalizations can be safely drawn. However, enoughis known to confirm again that bipolarity, in the sense of identicalspecies occurring in Arctic and Antarctic regions, does not existexcept in the case of ubiquitous or cosmopolitan species (particularlythe genus Colossendeis) which are found in deep water in all oceans.Fifty or sixty years ago the bipolar hypothesis received much atten-tion, but D'Arcy Thompson gave it a rough handling, pointing outthat the theory had been built upon a foundation of inadequate sys-tematics.-" Now the bipolar hypothesis, insofar as Arctic and Ant-arctic faunas is concerned, is no longer accepted, but the name lingerson and has been applied in a different sense than its originators in-tended. In the words of Sverdrup, Johnson, and Fleming, "bipolaranimals need not necessarily be bipolar." -*26 Far a comprehensive comparison between the Japanese and Californiacoasts, see Gislen, T., Physiographical and ecological investigations concerningthe littoral of the northern Pacific. Section I, A comparison between the lifeconditions in the littoral of central Japan and California, Univ. Arsskr. Lund,(2), vol. 39, No. 5, 63 pp., 1943, and Sections II-IV, Regional conditions of thePacific coast of America and their significance for the development of marinelife, ibid., vol. 40, No. 8, 91 pp., 1944.
^"^ On a supposed resemblance between the marine faunas of the Arctic andthe Antarctic regions. (Proc. Roy. Soc. Edinburgh, vol. 22, pp. 311-349, [1898].)28 Sverdrup, H. U., Johnson, Martin W., and Fleming, Richard H., Theoceans : Their physics, chemistry and general biology, p. 849, New York, 1942.
40 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL, I06To be sure, there are an extraordinarily large number of species ofNymphon in both Arctic and Antarctic waters, but the genus is farfrom rare in tropical waters, and all that can be certainly said of thisdistribution is that this genus flourishes best in the cold waters of thehigher latitudes. With one dubious exception {Nymphon longitarsevar. antarctictim) , there is no bipolar species of Nymphon. Anothersignificant fact is the absence of lo-legged forms from the Arctic,which gives a radically different facies to the Antarctic fauna.There are, however, some examples of distribution which may beconstrued as support of the revised bipolar—or abipolar—pattern ofdistribution. The most conspicuous example is that of the genusRhynchothorax, one species of which is known from the Isle of Capriin the Mediterranean, while the other appears to be a circumpolarAntarctic and Magellanic species. No intervening records are yetknown. Several genera, such as Achelia, Tanystylum, and Pallenopsis,prefer the temperate latitudes of both Northern and Southern Hemi-spheres and are poorly represented in the Tropics. The case of thebathypelagic Pallenopsis calcanea is not so clear. This species has beentaken at moderately great depths (600 to 800 fathoms) in DavisStrait and the Indian Ocean off South Africa. There is a third recordof this species from the vicinity of Bermuda, and it is possible thatthis may be a widely distributed species which has not often beencollected because of its bathypelagic habit.The Antarctic genus Austropallene is apparently the southerncounterpart of the Northern Hemisphere Cordylochele, but a SouthAfrican species described by Flynn (1928) as Pseudopallene gilchristidiffers from Cordylochele solely in the possession of a setose fringearound the mouth, and it is probably actually a Cordylochele.^^ Thiswould deprive Cordylochele of its status as a northern genus andweaken the "bipolar" relationship between Cordylochele andAustropallene.Exclusive of extra-legged forms, and of tropical genera which arefound on both sides of the Equator, there are perhaps 10 generaendemic to the Southern Hemisphere.^'' Of these, four are restricted
It is unfortunate that the archaic class designation Arachnoida, comprising seamites, pycnogonids, and Limultis, is sanctioned in this comprehensive treatise.29 The presence or absence of a setose fringe may be a specific character inthis genus as it appears to be in Pallenopsis. Cf. Pallenopsis denticulata Hedg-peth (1944).3° Ammothea s. str. may be a southern genus, but there are several dubiousNorthern Hemisphere records which are not yet confirmed. The taxonomicstatus of other genera is uncertain.
NO. l8 PYCNOGONIDA—HEDGPETH 4I
to the Antarctic (Austropallene, Austrodecus, Austroraptus, andHeteronymphon) , and two are known only from the Cape region ofSouth Africa {Boehmia, Ainigma). There are also two other generawhich are characteristic members of the Cape fauna (Discoarachneand Hannonia), but both of these have been identified from PortNatal. Pycnothea is so far known from one species at Juan Fer-nandez and another at Rottnest Island. The genus Oorhynchus isknown only from a deep-water species taken north of New Zealandby the Challenger. It will be noted that four of the Southern Hemi-sphere genera occur along the South African coast. This concentra-tion is not surprising in view of Ekman's (op. cit., p. 275) summaryof endemic forms from this region.Although there are many small genera in scattered parts of theTropics, there is only one large genus, Anoplodactyhis, which mightbe considered typically tropical. While it is represented by severalspecies in temperate latitudes, it attains its greatest speciation in theTropics, especially in the West Indies. There is also but one littoralgenus which might be said to be cosmopolitan in the sense that itsspecies occur in about the same proportions (usually two or threewell-differentiated species in any given region) throughout the oceans.This is Pycnogonuui, and its large number of endemic species is pos-sibly due to the heavy body form and sluggish movements which arecharacteristic of the genus.In general, it appears that the endemic genera of the SouthernHemisphere are more widely distributed than those of the Northern,which is not surprising in view of the more open character of thesouthern oceans. Each successive Antarctic expedition establishes thecircumpolar distribution of more Antarctic species, and littoral col-lections in the South Sea islands will doubtless bring to light manyspecies described from the East Indies and the Indian Ocean. A closerelationship between the fauna of South Africa and the East Indieshas already been noted by Flynn (1928, p. 3) who suggests that "thegreat equatorial current is responsible for such phenomena."SUMMARY1. The characters and ontogeny of the Pycnogonida entitle themto the stature of a class or subphylum of the Arthropoda, althoughtheir relationship to other groups of arthropods still remains uncertain.2. The Pycnogonida constitute a compact self-contained group offamilies without ordinal distinctions.3. The pattern of variation within the Pycnogonida is correlated
42 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. I06
with extra-legged or polymerous forms. In families where variationis most active at the generic rank, polymerous forms do not occur,whereas in families composed of a large number of species in one ortwo genera, several lo-, and in one case, 12-legged forms are pres-ent. The great majority of species, however, are octopodous.4. Extra-legged forms are closely similar to "normal" octopodousspecies, and may be polymorphic forms of these species. Certainlythe lo-legged forms and their cognate 8-legged species are representa-tives of the same racial stocks. The occurrence of this polymerouscondition in the Antarctic and tropical America, at the temperatureextremes of the marine environment, is suggestive of polyploidy asit occurs in many plants. These polymerous forms appear to be morenumerous and widely distributed than their cognate octopodous forms,suggesting that they are more successful from the evolutionarystandpoint.5. The Sargassum of the North Atlantic is an active agent in thedistribution of small, relatively immobile species in that ocean. Inthe North Pacific such small species tend to remain endemic.There is no evidence in the distribution of the Pycnogonida to sup-port the outworn concept of bipolarity. There are at least twiceas many endemic genera in the Southern Hemisphere, as contrastedwith northern waters, and most of these are widely distributed,whereas all the endemic genera (except an Atlantic deep-water formand an anomalous genus from Suez Canal) of the Northern Hemi-sphere are restricted to the Pacific, exclusive of tropical forms.BIBLIOGRAPHYThis bibliography comprises all the titles on Pycnogonida, insofar as I havebeen able to ascertain them, which have been published since the comprehensivebibliography in Heifer and Schlottke (1935), together with omissions fromthat work. References cited in this paper but not included here will be foundin the earlier bibliography. I wish to thank Dr. Isabella Gordon and Dr. ErnestoMarcus for assistance in making this compilation.Andre, M., and Lamy, E.1938. Pycnogonides parasites de moUusques. Journ. Conchyl., vol 82, pp.326-331-Arita, K.1936. Ein iiberziihliges Bein bei einer Pantopoden-Art (Nyinphonellatapetis Ohshima). Annot. Zool. Jap., vol. 15, No. 4, pp. 469-477,pi. 32, 3 figs.1937. Beitrage zur Biologic der Pantopoden. Journ. Dep. Agr., KyushuImp. Univ., vol. 5, No. 6, pp. 271-288, 7 figs.
NO. 1
8
PYCNOGONIDA—HEDGPETH 43BOUVIER, E. L.1937. fitude sur les Pycnogonides du "Travailleur" et du "Talisman"precedee d'observations systematiques sur les Articules de ce group.Ann. Sci. Nat, Zool. (10), vol. 20, No. i, pp. 1-42, 13 figs.Calman, W. T.1937a. James Eights, a pioneer Antarctic naturalist. Proc. Linn. Soc.London, 1936-1937, vol. 149, No. 4, pp. 171-184, pi. 4, 3 figs.1937b. The type specimens of Pallene atisfraliensis Hoek (Pycnogonida).Ann. and Mag. Nat. Hist. (10), vol. 20, pp. 530-534, 6 figs.1938. Pycnogonida. The John Murray Exp., Sci Rep., vol. 5, No. 6, pp.147-166, 10 figs.Carpenter, G. H.1903. On the relationships between the classes of Arthropoda. Proc. Roy.Irish Acad., vol. 24, No. B 4, pp. 320-360.1905. Notes on segmentation and phylogeny of the Arthropoda. Quart.Journ. Micr. Soc, vol. 49, No. 3, pp. 469-491.Dawson, A. B.1934. The colored corpuscles of the blood of the purple sea spider, Anoplo-dactylus lentiis Wilson. Biol. Bull., vol. 66, pp. 62-68, pi. I.Derugin, K. M. (editor).1935- Pantopoda of the Polar Seas within U.S.S.R. Inst. ArctiqueU.R.S.S., Materials for the Study of the Arctic, vol. 4, pp. 1-140,17 figs.EXLINE, H. I.1936. Pycnogonids from Puget Sound. Proc. U. S. Nat. Mus., vol. 83, No.2991, pp. 413-422, fig. 33-Page, L.1942. Pycnogonides de la cote occidentale d'Afrique. Arch. Zool. Exp. etGen., vol. 82, Notes et revue, pp. 75-90, 7 figs.Flynn, T. T.1929. Pycnogonida from the Queensland coast. Mem. Queensland Mus.,vol. 9, No. 3, pp. 252-260, 9 figs.GiLTAV, L.1934a. Pycnogonida from the coast of British Columbia. Can. Field Nat.,vol. 48, pp. 49-50, distributional list.1934b. Remarques sur le genre Ammothea Leach et description d'une especenouvelle de la mer d'Irlande. Bull. Mus. Roy. Hist. Nat. Belgique,vol. 10, No. 18, pp. 1-6, 3 figs.1934c. Note sur quelques Pycnogonides de Villefranche S/Mer (Alpes Mari-times). Ibid., No. 35, pp. 1-5, i fig.I934d. A new Pycnogonid from Bermuda. Ibid., No. 42, pp. 1-3, 6 figs.1935- Pycnogonides. Res. Voy. Belgica, Rap. Sci., Zool., pp. 3-16, figs. i-io.1937' Pycnogonida. Res. Sci. "Mercator," i. Mem. Mus. Roy. Hist. Nat.Belgique (2), vol. 9, pp. 83-89, 6 figs.1942. New records of Pycnogonida from the Canadian Atlantic coast.Journ. Fish. Res. Board Canada, vol. 5, No. 5, pp. 459-460.Gordon, I.1938. Pycnogonida. Australasian Antarctic Expedition, Sci. Rep. (C),Zool. and Bot., vol. 2, No. 8, pp. 1-40, 8 figs.1944. Pycnogonida. British, Australian, and New Zealand Antarctic Res.Exp., 1929-1931, Rep., ser. B, vol. 5, No. i, pp. 1-72, 27 figs.
44 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. Io6Hedgpeth, J. W.1939- Some pj^cnogonids found off the coast of Southern California. Amer.Midi. Nat., vol. 22, No. 2, pp. 458-475, 2 pis.1940. A new pycnogonid from Pescadero, Calif., and distributional notes onother species. Journ. Washington Acad. Sci., vol 30, No. 2, pp.84-87, I fig.1941a. On a new species of Nymphon from the waters of Southern Cali-fornia. Amer. Midi. Nat., vol. 25, No. 2, pp. 447-449, i pi.1941b. A key to the Pycnogonida of the Pacific coast of North America.Trans. San Diego Soc. Nat. Hist., vol. 9, No. 26, pp. 253-264,pis. 9-11.1942. Sea spiders. Nat. Mag., October, pp. 414-415, figs. (Popular article.)1943a. Pycnogonida of the Bartlett collections. Journ. Washington Acad.Sci., vol. 22, No. 3, pp. 83-90, 2 figs.1943b. Pycnogonida from the West Indies and South America collected bythe Atlantis and earlier expeditions. Proc. New England Zool.Club, vol. 22, pp. 41-58, pis. 8-10.1943c. On a species of pycnogonid from the North Pacific. Journ. Wash-ington Acad. Sci., vol. 22, No. 7, pp. 223-224, i fig.1944. On a new species of Pallenopsis (Pycnogonida) from western Aus-tralia. Proc. New England Zool. Club, vol. 22, pp. 55-58, pis. 11-12.(In press.) Pycnogonida. Encycl. Brit.The Pycnogonida of the western North Atlantic and the Caribbean.Proc. U. S. Nat. Mus.Report on the Pycnogonida collected by the Albatross in Japanesewaters in 1900 and 1906. Ibid.Helper, H.1935. Meerespinnen. Der Naturforscher, October. (Popular article.)1936a. The fishery grounds near Alexandria. VIII, Pantopoda. Ministry ofCommerce and Industry, Egypt, Notes and Mem., No. 16, pp. 1-6,3 figs.1936b. Pantopoda. (2 te Nachtrag.) Die Tierwelt Nord imd Ostsce, vol. 31,Teil XIa3, pp. 1-5, distributional map.1938. Einige neue Pantopoden aus der Sammlung des ZoologischensMuseums in Berlin. Sitzb. Ges. Nat. Fr. Berlin, 1937, pt. 2, pp.162-185, II figs.Helper, H., and Schlottke, Egon.1935- Pantopoda. Bronns Kl. u. Ordn. Tierreichs, vol. 5, Abt. IV, Buch 2,Lief. I, pp. 1-160, 223 figs.Hilton, W. A.1934. Notes on parasitic pycnognids [sic]. Pomona Journ. Ent. and Zool.,vol. 26, No. 4, p. 57.1939a. A preliminary list of pycnognids [sic] from the shores of California.Ibid., vol. 31, No. 2, pp. 27-35.1939b. A collection of pycnognids [sic] from Santa Cruz Island. Ibid.,No. 4, pp. 72-74, 1 1 figs.1942a. Pantopoda. Pantopoda [sic] chiefly from the Pacific. I. Nympho-nidae. Ibid., vol. 34, No. i, pp. 2-7-1942b. Pycnogonids from Allan Hancock Expeditions. Allan Hancock Pa-cific Exp., vol. 5, No. 9, pp. 277-312, pis. 35-48.
NO. 1
8
PYCNOGONIDA—HEDGPETH 45
1942c. Pantopoda (continued). II. Family Callipallenidae. Pomona Journ.Ent. and Zool., vol. 34, No. 2, pp. 38-41.I942d. Pycnogonids from Hawaii. Occ. Pap. Bishop Mus., vol. 17, No. 3,pp. 43-55, 10 figs.19426. Pycnogonids from the Pacific. Family Tanystylidae. Pomona Journ.Ent. and Zool., vol. 34, No. 3, pp. 69-70.I942f. Pycnogonids from the Pacific. Family Phoxichilididae [sic] Sars1891. Ibid., pp. 71-74.1943a. Pycnogonids from the Pacific. Family Ammotheidae. Ibid., vol. 34,No. 4, pp. 93-99.1943b. Pycnogonids from the Pacific. Family Colossendeidae. Ibid., vol. 35,No. I, pp. 2-4.1943c. Pycnogonids of the Pacific. Family Pycnogonidae. Ibid., No. 2, p. 19.Lebour, M. V.1945. Notes on the Pycnogonida of Plymouth. Journ. Mar. Biol. Assoc.U. K., vol. 26, pp. 139-165, 7 figs.LOMAN, J. C. C.1938. Note preliminaire sur les "Podosomata" (Pycnogonides) du MuseeOceanographique de Monaco. Res. Camp. Sci. Monaco, vol. 97,pp. 277-286, pi. 4, figs. II, 19-24. (Reprint of Loman, 1912.)Lou, Ting-Heng.1936. Sur deux nouvelles varietes des pycnogonides receuilles a Tsing-Tao,dans la baie de Kiao-Chow, Chine. Contr. Inst. Zool., Nat. Acad.Peiping, vol. 3, No. i, pp. 1-34, 4 pis., 9 figs.Mane-Garzon, F.1944. Notas sobre pantopodos. I. Colossendeis geoffroyi, nov. sp., de laplataforma continental frente al Rio de la Plata. Comm. Zool. Mus.Hist. Nat. Montevideo, vol. i, No. 15, pp. 1-7, 4 figs.Marcus, E.1940a. Pallenopsis fluminensis (Kroyer) e as Pallenopsis sul-atlanticasrestantes. Rev. Ent. Rio de Janeiro, vol. 11, Nos. 1-2, pp. 180-199,I fig.1940b. Os Pantopoda brasileiros e os demais sul-americanos. Bol. Fac. Fil.,Cien. e Letr. Univ. Sao Paulo, vol. 19, Zool. 4, pp. 3-179, pi. 1-17.194OC. Os Pantopoda. Fil. Cien. e Letr., Orgao do Gremio do Fac. Fil.Cien. e Letr. Sao Paulo, vol. 7, pp. 68-73. (Popular summary.)Mello-Leitao, A. DE.1945. Uma especie nova do genero Pycnogonum Briinnich, 1764. (Pycno-gonidae, Pantopoda.) Bol. Mus. Nac. Rio de Janeiro, n.s., Zool.,No. 42, pp. 1-4, 5 figs.Moore, H. B.1933. New faunistic records for the Manx region. Rep. Mar. Biol. Stat.Port Erin, vol. 46, pp. 30-34.1934. New faunistic records for the Manx region. Ibid., vol 47, pp. 27-28.Morley, C.1940. Arachnida of Suffolk: The remaining Arachnida. Trans. SuffolkNat. Soc, vol. 4, No. 3, pp. 165-174.Needler, a. B.1943. Pantopoda. Can. Atlantic Fauna, vol. 10. Arthropoda, ion, Panto-poda, pp. 1-16, 21 figs.
46 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. I06Ohshima, H.1935a. "On a sea spider inhabiting the Okinawa region." DobutsugakuZasshi (Zool. Soc. Japan), vol. 47, pp. 137-139. (In Japanese.)1935b. A further note on Nymphonella tapetis : The egg-carrying maturemale. Annot. Zool. Jap., vol. 15, No. i, pp. 95-102, 4 figs.1936. A list of Pycnogonida recorded from Japanese and adjacent waters.Zool. Mag., vol. 48, Nos. 8, 9, 10, pp. 861-869.1937. The life-history of "Nymphonella tapetis" Ohshima. Compt. Rend.XIP Congr. Internat. Zool., Lisbonne, 1935, pp. 1616-1625, pi. 80,5 figs.1938. Nymphonellidae, a new family of Pantopoda. Annot. Zool. Jap., vol.17, Nos. 3, 4, pp. 229-233.1942a. Six-legged Pantopod, an extraordinary case of hypomery in arthro-pods. Proc. Imp. Acad. Tokyo, vol. 18, No. 5, pp. 257-262, 3 figs,1942b. A remarkable case of malformed appendages in a pantopod, Nympho-nella tapetis. Ibid., No. 8, pp. 520-523, 2 figs.1943a. [Resume of 1942b in Japanese.] Rep. Ann. Meet. Zool. Soc. Japan1943, I P- (No page numbers in reprint.)1943b. Du maloftaj kazoj de nenormaleco en pantopodoj. Bui. Sci. Fak.Terk. Kyusu Imp. Univ., vol. 10, No. 4, pp. 371-382, 4 figs. (Japa-nese version of 1942a and 1942b, with short Esperanto summary.)Pelt, W. G.1936. Pantopoda. Flora en Fauna der Zuiderzee, suppl., pp. I3i-i33-Sawaya, M. p.1941. Sobre uma larva de Pycnogonuin pamphornm Marc. Bol. Fac. Fil.,Cien. e Letr. Univ. Sao Paulo, vol. 22, Zool. 5, pp. 278-282, 2 figs.1945. Anoplodactyhts stictus Marc. (Pantopoda) em Caioba, Estado doParana. Arq. Mus. Paranaense, vol. 4, No. 10, pp. 231-236, pi. 23.SCHMITT, W. L.1934. Notes on certain pycnogonids including descriptions of two newspecies of Pycnogonuin. Journ. Washington Acad. Sci., vol 24,No. I, pp. 61-70, 2 figs.1945. Miscellaneous zoological material collected by the United StatesAntarctic Service Expedition, 1939-1941. Proc. Amer. Philos. Soc.,vol. 89, No. I, p. 297.Snodgrass, R. E.1938. Evolution of the Annelida, Onychophora, and Arthropoda. Smith-sonian Misc. Coll., vol. 97, No. 6, pp. 1-159, 54 figs.Stephensen, K.1936a. Sveriges pycnogonider. Kungl. Vetensk. Vitt.-Sam. Handl. Goteborg(B), vol. 4, No. 14, pp. 1-56, 13 figs. (Meddl. Goteborgs Mus. Zool.Avd., vol. 69.)1936b. Pycnogonida from Norway and adjacent waters. Bergens Mus. Arb.1935, No. 7, pp. 1-39, I fig.1937. Pycnogonida. Zool. Iceland, vol. 3, No. 58, pp. 1-13.1938. Amphipoda. Tanaidacea and Pycnogonida. Zool. Ergebn. Reis. Dr.Kohl Larsen . . . ., vol. 20, Nos. 3-4, pp. 236-264.1943. Pycnogonida. The Zoology of East Greenland. Meddl. Gronl., vol.121, No. 8, pp. 1-41, 7 figs.
NO. 1
8
PYCNOGONIDA—HEDGPETH 47
TOPSENT, E.1938a. Les pycnogonides provenant des campagnes du yacht VHiroiidelle,1886-1887-1888. (Golfe de Gascogne, Terre-neuve, Agores.) Res.Camp. Sci. Monaco, vol. 97, pp. 272-276. (Reprint of Topsent,1891.)1938b. Pycnogonides receuillis par le yacht Princesse Alice. Ibid., pp. 276-277.(Reprint of Topsent, 1897.)Williams, G.1933- On Nymphopsis acinacispinatns, a new pycnogonid from Queensland.Ann. and Mag. Nat. Hist. (10), vol. 12, pp. 173-180, 6 figs.1940. Pycnogonida of Western Australia. Journ. Roy. Soc. Western Aus-tralia, vol. 25 (1938-1939). PP- 197-205, 9 figs.1941. A revision of the genus Anoplodactyliis with a new species fromQueensland. Mem. Queensland Mus., vol. 12, No. 1, pp. 33-39.5 figs.
48 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. I06
NO. 1
8
PYCNOGONIDA HEDGPETH 49
Qagou
IQZw
50 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. I06
NO. U PYCNOGONIDA—HEDGPETH 51
— g
^1
52 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. Io6
1
xo. i8 PYCNOGONIDA—HEDGPETH 53
z§
ft.
-;->

SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 106, NO. 18, PL. 1
PycnogonidaA, Neoliolotype, Ih-colopoda tiiistralis Eights; B, Fciitacoh'ssc'iidcisreticulata Hed'gpcth ; C, Fcntauymphou antarcticum Hodgsdii. (A,about 2 X ; B. nearly natural size; C. slightly enlarged.)