maxmen amy 2008

8/6/2019 Maxmen Amy 2008

1/13

The Sea Spiders Contribution to T.H. Morgans

(18661945) Development AMY MAXMENDepartment of Organismic and Evolutionary Biology and Museum of Compara-tive Zoology, Harvard University, Cambridge, Massachussets

ABSTRACT Over a century ago, T.H. Morgan helped found The Journal of ExperimentalZoo logy, a series devoted to emerging investigations in development and evolution, variation andheredity, adaptation and ecology. T.H. Morgan initially encountered these topics in his graduateresearch on the ontogeny and phylogeny of the sea spiders. His Ph.D. thesis, written in 1891, reflectsin interesting ways the conceptual shifts in evolutionary biology in the late 1800s. Embryology hadbecome a major criterion in using morphological similarity to speculate on phylogenetic relation-ships. Yet, when Morgan studied the development of sea spiders to draw conclusions about their

relatedness, he struggled with the incompleteness of knowledge about the role of inheritance andvariation of ontogenetic processes. This can best be seen in his discussions of the properties ofconservation during embryonic cleavage, varied development of supposedly homologous appendages,and the evolvability of larval stages. After his dissertation, Morgan never returned to phylogeneticanalysis. He had been dissatisfied with the plethora of untestable phylogenetic hypotheses based oncomparing complicated embryological phenomenon, and joined an experimental movement in whichsystems were explicitly chosen and constructed to test hypotheses about developmental processes.

J. Exp. Zool. (Mol. Dev. Evol.) 310B:203215, 2008. r 2007 Wiley-Liss, Inc.

How to cite this article: Maxmen A. 2008. The sea spiders contribution to T.H. Morgans(18661945) development. J. Exp. Zool. (Mol. Dev. Evol.) 310B:203215.

In the summer of 1890, Thomas Hunt Morgan(Fig. 1), age 24 years, delivered a lecture at theMarine Biological Laboratory at Woods Hole,Massachusetts, based on his dissertation research.Woods Hole was emerging as a center for scientificthoughts in the 1890s, and Morgan was in goodcompany. His advisor, William Kenneth Brooks(18481908), a classical morphologist, was respon-sible for mentoring future cytologist, Edmund B. Wilson (18561908), embryologists Ross G.Harrison (18701959) and Edwin G. Conklin(18631952), and geneticist William Bateson(18611926) among others at Johns Hopkins(Maienschein, 91b). Brooks had, in turn, been astudent of Louis Agassiz (18071873) and later Alexander Agassiz (18351910) at HarvardsMuseum of Comparative Zoology. Brooks carriedon his mentors precise methods of descriptivemorphology, without endorsing Louis Agassizcreationist agenda. He was well known forencouraging his students to think independentlyand he did not expect that answers about evolu-tionary relatedness drawn from comparative zool-ogy should come easily. Bateson (10) would later

say of his mentor, Variation and heredity hadstood as axioms. For Brooks they were problems. As he talked of them the insistence of theseproblems became imminent and oppressive.

Morgans dissertation on the evolutionary rela-tionship of sea spiders (pycnogonids, the extantmembers of Pantapoda) among arthropodscontains one of the few embryological accounts ofpycnogonid development to this day. Yet, exclud-ing a handful of sea spider specialists, few arefamiliar with the work, in part because Morgansdissertation is an unexceptional outlier among hisapproximately 370 scientific manuscripts and 22books. Using comparative anatomy and embryol-ogy to infer phylogeny was the traditionalapproach for an evolutionist in 1890, and thesestudies contrast to Morgans later and more

Published online 30 October 2007 in Wiley InterScience (www.interscience.wiley.com). DOI: 10.1002/jez.b.21202

Received 5 September 2007; Accepted 14 September 2007

Correspondence to: Amy Maxmen, Department of Organismic andEvolutionary Biology and Museum of Comparative Zoology, HarvardUniversity, 26 Oxford St., Cambridge, MA 02138.E-mail: [email protected]

r 2007 WILEY-LISS, INC.

JOURNAL OF EXPERIMENTAL ZOOLOGY (MOL DEV EVOL) 310B:203215 (2008)

8/6/2019 Maxmen Amy 2008

2/13

famous experimental work for that he would win aNobel Prize in 1933. His descriptive graduatework has been laid to rest in dimly lit librarystacks and when it is acknowledged, it is givenonly a cursory analysis. For instance, the geneti-cist Sturtevant (59), a former student of Morgan,and the historian of science, Maienschein (91b)considered the thesis merely typical of a graduatestudent studying zoology in the late 19th century.Historian of science, Allen (69, 78) discussed thedissertation in his biography of Morgan, butcertain important details are obscured.

In this study, I critically evaluate Morgansthree manuscripts on sea spider phylogeny: A Preliminary Note on the Embryology of thePycnogonids (1889), The Relationships of theSea Spiders (1890), and his dissertation and A Contribution to the Embryology and Phylogenyof the Pycnogonids (1891). I argue that Morgan didnot merely pass through the prescribed andschematic motions of using ontogenetic stages

as evidence for relatedness; rather, Morganstruggles with the inevitable incompletenessof embryological knowledge culminating in anexplicit skepticism about his own phylogeneticconclusions.

The manuscripts reflect a growing discontentwith using embryology to deduce evolutionaryrelationships. In Morgans (1889) first scientificpublication, he is hopeful that embryologicaldescriptions will reveal the phylogenetic relation-ships of the pycnogonids among arthropods. InMorgans (1891) thesis this initial optimism hasdisappeared. In the thesis, most hypotheticalevolutionary scenarios are removed and the topicsmore often than not include open-ended questionsrather than ready answers. In some cases thesechanges in Morgans writings are subtle. As historian of embryology, Oppenheimer (67)

observed, It is one of the obstacles to an approachto certainty in intellectual history thatthe influence of one idea or pattern of ideas onthe development of another is not indicatedaccurately by references either in texts or biblio-graphies (p 13). Here, I pay particular attentionto the tone, questions, and discrepancies withinand between Morgans graduate manuscriptsto discover why Morgan was initially interestedin sea spiders and comparative embryology, andwhy he eventually found his own results to beproblematic.

I focus on those aspects of Morgans graduateinvestigations that are relevant for todays discus-sions. Many of the problems Morgan grappledwith predate the vocabulary we now use todescribe them,1 namely, process (or transforma-tional) homology, homoplasy, and constraint (withrespect to an ancestral larval or phylotypicstage). These topics implicitly arise in hisdiscussion of embryonic cleavage, germ layerformation, appendicular homology, and the evol-vability of the sea spider larva, respectively. Onecentral problem for Morgans interpretationsinvolved the supposed polarization of concepts ofheredity and variation in 1880. Morgan wasfamiliar with Francis M. Balfours comprehensiveTreatise on Comparative Embryology,2 in whichheredity and variation are presented as interact-ing dialectical opposites. There are, accordingto this theory, two guiding, and in a certain

Fig. 1. A photograph of Thomas Hunt Morgan from thefrontispiece of Journal of Experimental Zoology (1945).Morgan published a manuscript in this volume 100, as wellas in the first (1904).

1 Morgan rarely introduced new terms and discouraged thehabit. While giving the appearance of profundity, in reality oftenserve merely to cover ignorance and to make mystery out of amechanism (Morgan, 19; p 59).

2 For evidence that Morgan was familiar with Balfour, pleaserefer to the discussion in this manuscript on Larval Evolution.

A. MAXMEN204

J. Exp. Zool. (Mol. Dev. Evol.)

8/6/2019 Maxmen Amy 2008

3/13

sense, antagonistic principles that have renderedpossible the present order of the organic world.These are known as the laws of heredity andvariation (Balfour, 1880; p 2). The theory seemedto be straightforward but Morgan observed

unaccountable variation at different points alonga developmental trajectory in related species thatmade him question some of the fundamentalassumptions about heredity and development.

PHYLOGENETIC POSITION OF SEASPIDERS

It is quite understandable why Morgan wasinterested in the open challenge of sea spiderphylogeny. Pycnogonids are bizarre arthropods,characterized by bodies too slender to house thegut diverticula and gonads, which are instead

located in the four to six pairs of long legs. Thehead typically bears a snout-like proboscis, and apair of clawed appendages (chelifores), palps, andspecialized appendages (ovigers), which the maleuses to carry embryos until hatching (Fig. 2A).They are exclusively marine, cosmopolitan, andmost importantly for Morgan, certain species havebeen making their home on alga and hydroid beds

nearby Woods Hole, Massachusetts, for over acentury (Fig. 2B).

In the decade before Morgans own graduatework, Hoek (1881; affiliated with the ChallengerExpeditions), Dohrn (1881), director of the marine

laboratory, Stazione Zoologica at Naples and astudent of Ernst Haeckel, and Edmund B. Wilson(1878, 1881; Morgans classmate at Johns Hopkinsand a future leading cytologist) had all publishedmanuscripts on sea spider evolution. Wilsonspycnogonids were collected by Alexander Agassizin 1880. It is quite possible that Agassiz pycno-gonids had been given to Wilson by Brooks andlater Wilson might have impressed Morgan withtheir potential for further phylogenetic analysis.Although all these scientists had put forth phylo-genetic theories, the actual hypotheses changedfrom manuscript to manuscript.

But Pycnogonids were also an important taxonto understand. Sea spiders were considered pri-mitive and thus they were paramount in under-standing the origin and evolution of the Arthropoda. Sea spiders affinity had affiliatedwith both crustaceans and arachnids and, in 1881,Dohrn and Hoek independently decided theywere an entirely separate lineage of arthropods,

Fig. 2. (A) Figure from Morgan (1890) depicting an adult sea spider, or pycnogonid. (B) A sea spider, Tanystylum orbiculareWilson, on a sargassum bladder in Woods Hole, Massachusetts (Hedgepeth, 48).

T.H. MORGAN AND THE SEA SPIDERS 205


8/6/2019 Maxmen Amy 2008

4/13

descendants of a common arthropodannelid an-cestor. On the basis of comparative studies of theiranatomy, Wilson (1881) concluded that the firstpair of appendages were significantly differentbetween arachnids and pycnogonids, and used this

character to separate the two groups. In CarlGegenbaurs (1878) Elements of Comparative Anat-omy pycnogonids were grouped beside anotherbizarre invertebrate, the water-bear or tardigrade,as an unsettled subgroup of arthropods. And in ATreatise on Comparative Embryology, Balfour(1880) suggested that pycnogonids are affiliatedneither with crustaceans nor arachnids, andthat they must have diverged very early froman arthropod ancestor. He recommends furtherinvestigation into pycnogonid embryology andappendicular innervation for future comparativestudies.

EMBRYOLOGY IN PHYLOGENY

Morgan (1890) begins his lecture, The Relation- ships of the Sea Spiders by announcing that hispoint of attack is to be largely from the sideof the embryology of the groups (p 145).E.B. Wilson had emphasized adult sea spideranatomy, A. Hoek and P.P.C. Dohrn emphasizedthe larval phase, but both had neglected pycnogo-nid embryology. Within the theoretical frame-work of evolutionary morphology comparative

embryology was considered imperative fortracing the evolutionary history of differentgroups, offering a criterion independent fromcomparative adult anatomy for determining rela-tionships. Early in the 18th century, the Germananatomist, Johann Friedrich Meckel (17811833)put forward the theory that developmental stagesrevealed the progression of the organism fromlower forms to its present state. Throughout the19th century, morphologists acknowledged thesimilarity between embryonic stages among mem-bers of a group and promoted the ideas from theGerman embryologist Karl Ernst von Baer(17921876). Von Baer maintained that the ten-dency of change during development was from theundifferentiated and general to the specific anddetermined. In this way, likeness between animalsduring early ontogeny could be used as evidencefor phylogenetic relationships. Furthermore, Dar-wins theory of evolution by natural selection hadprovided an explanation for why earlier stagesmight be expected to be conserved across taxalargely immobile and nonfeeding early stages didnot seem to competitively interact with their

environment and thus the struggle to adaptdiminished.

Thus, if Morgan was interested in finding achallenging thesis project, adding comparativeembryology to his pursuit of sea spider phylo-

geny would have been recommended as onesuch topic. In A Treatise on Comparative Embry-ology, Balfour notes that if early stages clearlyfollowed the conserving laws of heredity aspredicted by Darwin and Von Baer, phylogeneticreconstruction would be simple. He warns, how-ever, that the embryological record, as it isusually presented to us, is both imperfect andmisleading (Balfour, 1880; p 3). E.B. Wilson hadnoticed as well that trends in development andevolution were not simple to detect.

Every living being, at every period of itsexistence, presents us with a double problem.First, it is a complicated piece of mechanism,which so operates as to maintain, actively orpassively, a moving equilibrium between itsown parts and with its environment .... But inthe second place, the particular character ofthis adaptation cannot be explained by refer-ence to existing conditions alone, since theorganism is a product of the past as well as ofthe present ... Phenomena of the latter classmay, for the sake of brevity, conveniently betermed ancestral reminiscences (Wilson,

1878; p 1).

CLEAVAGE AND GERM LAYERFORMATION

Complications aside, Morgan began his work byobserving and documenting the development ofthree pycnogonid species. In general, Morgancompared embryological processes rather thansingle structures during development, reflectingthe fact that the developing embryo is hardlystatic, and therefore difficult to compare at asingle point in time.

The initial series of mitotic cell divisions follow-ing the egg stage was called the segmentation orcleavage of the early embryo. Around 1890, celllineage studies were still in their infancy, and soinstead, it was common practice to compare thesegmentation patterns across taxa. Morgansproblems began when he observed variationin segmentation between pycnogonid species.Outside of the variation attributable to yolkquantity there were no sound theories for

A. MAXMEN206


8/6/2019 Maxmen Amy 2008

5/13

why one egg would develop differently fromanother. There seemed no adaptive advantagefor variation and the species were all clearlyderived from a common pycnogonid ancestor, thuspossessing the same heritable substance.

Morgan (1891) suggested that mechanical forces,such as gravity and pressure, as opposed toheritable programs, could be factors during devel-opment, but available data could not discriminateamong these possibilities:

The embryo differentiates earlier in whatcorresponds to the anterior region of the adultthan over the whole ventral surface, whichsuggests that the smaller cells may haveadapted themselves to this early differentia-tion; but it seems equally possible that this

differentiation may be due to phylogeneticlaws in this particular case rather than to anymechanical connection with the micromeredifferentiation. So that for the present thequestion must remain unsettled until byactual experiment (which would not be diffi-cult) the orientation of the segmentationplanes be determined (p 22).

To observe variation at this early stage indevelopment must not have been a completesurprise to Morgan. Other embryologists had

already questioned the reliability of segmentationin divulging relatedness (Balfour, 1880). Althoughsome cleavage patterns, such as spiral cleavage inwhich the orientation of cell spindles at mitosis isat a determined angle relative to the egg axis, heldtrue for most mollusks and annelids, other groups,such as the crustaceans, cleaved irregularly.Balfour (1880) admitted that ... similarity ordissimilarity of segmentation is no safe guide toaffinities. In many cases, it is true, a special type ofsegmentation may characterize a whole group; butin other cases very closely allied animals presentthe greatest differences with respect to theirsegmentation (p 100). The particles of inheri-tance were a black box and it was not clear if thepattern of early cell division was a consequence ofinherited material or if the characteristic featuresof the organism were resting in the initial cells aspotential only to be expressed later in thedevelopment.

The primary embryological phenomenon thatMorgan emphasizes in his studies is the formationof endoderm. Both in Balfours (1880) A Treatise on Comparative Embryology and Gegenbaurs

(1878) Elements of Comparative Anatomy, theformation of germ layers was considered phylo-genetically important as opposed to the lessunderstood and more variable process of embryo-nic segmentation. Organs could be observed

forming from germinal layers, and therefore theselayers were at the root of structural homologywhen anatomical comparisons were made betweentaxa. As a result of decades of emphasis on germlayer formation, Balfours (1880, 1885) Treatise,includes a collection of case studies, summarizingwhat was known about germ layer formationacross taxa. Two main processes that weredistinguished at this stage are invagination(Fig. 3A), in which cells ingress at a single pointto form the endoderm, and delamination, in whichinner cells divide to form the endodermal layerand outer cells become ectodermal. Morgan

describes the formation of pycnogonid germ layersas multipolar delamination (Fig. 3B). Duringmultipolar delamination the sphere of cellscomprising the early embryo divides simulta-neously, and inner cells give rise to earlyendoderm. The process Morgan observed inpycnogonids appeared to be similar to what hadbeen described previously in false-scorpionsand spiders.3 Multipolar delamination was treatedas an important similarity (process homologyin modern terminology) uniting arachnids andpycnogonids to the exclusion of crustaceans.

In his dissertation, Morgan (1891) expanded thediscussion of multipolar delamination to includethe alternative interpretation that such a processmay be acquired independently. He notes thatinsects have been described as forming endodermsimilarly, yet insects share little else with thepycnogonids that might indicate special affinity,or rather, descent from a recent common ancestor.Morgan, however, was not the first to question theevolutionary significance of the germ layerprocess. Simultaneously, Haeckels contemporary,the German anatomist Wilhelm His (18311904),was attacking germ-layer homology by demon-strating that the causes of germ-layer movementsmight be governed by the mechanics of cell growth(reviewed in Maienschein, 91a).

Even though these commentaries from Morgansdissertation might seem to be subtle and disjunc-tive, at the time they posed substantial questionsthat would prove consequential in the future.

3 These reports refer to Metschnikoff, E. (Entwicklungs-geschichte des Chelifer. Zeit. f. wiss. Zool., XXI, 1871) and Balfour(Notes on the Development of the Araneina. Quart J. Micr. Sci.,vol. XX, 1880).



8/6/2019 Maxmen Amy 2008

6/13

Uncertainty regarding ontogenetic flexibilityshook the foundations of contemporary phyloge-netics. Without understanding the mechanismbehind cell specification and differentiation duringmultipolar delamination, the extent to whichthese embryonic processes were predeterminedand heritable was speculative at best.4 Morganwas undoubtedly aware of the movement towardexperimental embryology called Entwicklungsme-chanik (developmental mechanics) spearheaded

by Wilhelm Roux (18501924) and others inEurope, who sought to understand the relativeimportance of external and internal factors indevelopment. The basic premise of Entwicklungs-mechanik was that embryonic development wasdifficult to understand simply by observation andthat experimentation provided a way to answerquestions about what was occurring within andbetween cells. Soon after his graduation, Morganwould travel to the Stazione Zoologica in Naples toinvestigate the role of the cytoplasm in thetransfer of cellular information during develop-ment and regeneration. Morgan (1898)became oneof the movements leading advocates in the US,explaining its goals to a wider audience, ... in thedefinition of developmental mechanics as thestudy of the causal morphology of the organism,Roux means simply that the changes in formthrough which the embryo passes are the result ofa series of causes, and these causes are what thenew study proposes to investigate (p 12).

HOMOLOGY OF APPENDAGES

Most studies in arthropod evolution discusswhich appendages might be equivalent amongtaxa (Table 1). In Elements of ComparativeAnatomy, Gegenbaur (1878) emphasizes the goalof using phylogeny to reveal anatomical homologyand to infer adaptation with certainty based onvariation between homologous structures (p 9). AsPycnogonid limbs are structurally dissimilar to

the limbs of other arthropods the job of assigningappendicular homology is particularly compli-cated, and, to this day, remains in flux (Buddand Telford, 2005; Dunlop and Arango, 2005;Maxmen et al., 2005; Jager et al., 2006). Morganregarded the problem off appendicular homologyin the pycnogonids as the stumbling block,repeating the phrase that Wilson (1881) had usedearlier. On the basis of embryology, Morgan wasuniting the pycnogonids with the arachnids, andfollowing this hypothesis he aligned the appen-dages between the two groups. Initially it wouldappear that the four pairs of pycnogonid walking

legs5 correspond to the four pairs of legs observedin arachnids. Pycnogonids, however, also havethree appendages associated with head segments(chelifores, palps, and ovigers), whereas arachnidshave two (chelicerae and pedipalps). In his earlywork, Morgan (1890) assumed that pycnogonidand spider appendages correspond alongthe length of the body, and that the extra

Fig. 3. (A) Figure modified from Balfour (1885) depicting an optical section during germ layer formation by the process ofinvagination in the sea cucumber, Holothuria tubulosa. (B) Figure from Morgan (1891) depicting a section of a sea spider(Tanystylum) embryo during germ layer formation by the process of multipolar delamination.

4 For an evaluation of the role of experimental embryology inthe emergence of the chromosomal theory of inheritance refer toGilbert (78).

5 Some pycnogonids bear five or six pairs of appendages,complicating the problem further.

A. MAXMEN208


8/6/2019 Maxmen Amy 2008

7/13

appendage is the final pair of pycnogonid walkinglegs, which, he explains, have been subsequentlylost in other arachnids. To account for theappendicular transformation in the evolution ofthe lineages from a pycnogonidarachnid ancestor,Morgan (1890) presents a hypothetical scenario,We may imagine, if we like, that this took place ata time when the third pair of appendages [theovigerous legs] appeared and began to carry theeggs, so that the body, by utilizing the first legs asegg carriers, retained a pair of the abdominal legsfor purposes of locomotion (p 156).

In contrast, hypothetical scenarios are conspicu-ously absent from Morgans dissertation. Morgan

(05) would later denounce such unsubstantiatedstories, It seems to me that the method of theDarwinian school of looking upon each particularfunction or structure of the individual as capableof indefinite control through selection is funda-mentally wrong (p 57). Morgans change inattitude can be seen when comparing his spec-ulative scenario from 1890 on leg evolutionbetween sea spiders and arachnids with a laterstatement on scientific methodology, [Hypoth-eses] have been used ... to hold together a body ofisolated facts; as such they are in reality onlyfictions. They have been used in the reconstruc-tion of supposed historical events, especially in

TABLE 1. The homologies of pycnogonid appendages to those of other arthropods alluded zoologists for over a centuryy

yWilson (1881) and Morgan (1891) referred to the problem as the stumbling block in pycnogonid phylogenetics. Table 1 is reproduced from

Hedgepeth (54: p 200) and figures of the pycnogonid, horseshoe crab (to represent Xiphosura), and spider (to represent arachnid) have been

added by the author.



8/6/2019 Maxmen Amy 2008

8/13

biology in the setting up of family trees (Morgan,07; p 6).

A corollary of Morgans hypothesis that appen-dages of pycnogonids and arachnids were homo-logous along the length of the body was that the

first pair of appendages, pycnogonid chelifores andarachnid chelicerae, were homologous. Specialhomology, as defined by Gegenbaur (1878), re-quired that the two structures be of similar descentand also develop from the same anlagen. For thisreason, Wilson (1881) and Dohrn (1881) hadrejected the homology of pycnogonid cheliforesand arachnid chelicerae because of the differencein the origin of nerves targeting the first appen-dages. Balfour (1880) referred to their studies whenhe distinguished the pycnogonids, with anteriorappendages innervated from the supraoesophagealganglion, from horseshoe crabs whose appendages

are all post-oral. In the early works, Morgan (1889,1890) ignores the voices of dissent and supportscorrespondence of chelifores and chelicerae withsimple, unillustrated descriptions of structuralsimilarity, innervation from the brain, and acommon developmental originboth appendagesarise heterolateral to the stomodeum. The conclu-sion remains the same in his dissertation, yet thediscussion is expanded to include ontogeneticdivergence in the formation of the first appendages(Fig. 4). According to Gegenbaur (1878; p 64),assigning special homology was the ultimate task

of comparative anatomy, and it required corre-sponding structures to share a common origin aswell as common descent. Morgan (1891) hadalready concluded that pycnogonids were relatedto arachnids based on embryological processes, andalthough he agreed that during developmentarachnid appendages are innervated at a moreposterior point than in pycnogonids, he maintainedthat the two appendages remained homologs,Chelate appendages are innervated in the adults[pycnogonids] by supraoesophageal gang-lia ... though, all the appendages in the embryos oftrue arachnida are innervated by post-oral gang-lia ...the early innervation of this pair from thebrain itself may be regarded as a more abbreviatedcondition than what was seen in spiders (p 30).Disparity during neurogenesis is noted, Theabsence of brain invaginations would be a moreweighty objection against the relationship of thetwo groups [arachnids and pycnogonids] (Morgan,1891; p 30). Ultimately, however, the disparity isdismissed, in that Morgan does not take thisvariation as necessitating reassessment of homol-ogy statements.

Likewise, Morgan (1891) describes and subse-quently downplays a potentially important featureof neurogenesis relating pycnogonids to the ony-chophoran (velvet worm), Peripatus. Compar-

ing ventral organ development in thepycnogonids with Peripatus,6 a striking similar-ity is seen ... Whether these structures are in anyway related is impossible to say, but it isworthwhile to call attention to the close similarityboth in position and structure between theseorgans in the two groups (p 24). Onychophoranswere considered to be potential intermediatesbetween annelids and arthropods, thus this char-acter would support a separate lineage of pycno-gonids. Yet, Morgan refrains from suggesting thatthe similarity in neurogenesis between pycnogo-nids and onychophans should override similarityof germ layer formation between pycnogonids andarachnids. There was no clear reason to favor onecriterion over another.

In retrospect, a confounding problem was thatphylogenetic methods were often circular prior towork by the systematist Willi Hennig (19131976).Homologous traits were defined by phylogeny, andthose traits were then recycled for reconstructing

Fig. 4. Figure from Morgan (1891; Plate I, Fig. 14)depicting a section through the anterior portion of thepycnogonid larva ofTanystylum orbiculare. The section passesthrough the brain (B), the circumoesophageal ring, and a pairof ventral ganglia (V). On either side of the brain are a pair ofdiverticula (D, 0D) leading to chelifores that receive nervesfrom the brain.

6 Morgan is referring to Sedgwick, A. (A Monograph of theDevelopment ofPeripatus Capensis. Stud. Morph. Lab., Camb., IV,Pt.I, 1888).

A. MAXMEN210


8/6/2019 Maxmen Amy 2008

9/13

phylogenies. Morgans comparative embryologyhad convinced him of a relationship betweenpycnogonids and arachnids, and so appendages ofpycnogonids and arachnids simply had to behomologous. Not understanding the extent to

which developmental trajectories were flexibleundermined confidence in homology statementsand this was a consistent source of conflict inphylogenetic reconstruction.

LARVAL EVOLUTION

A large portion of Morgans attention wasdevoted toward establishing the relevance ofthe post-embryonic larval stage, the protonym-phon, in phylogenetic reconstruction (Fig. 5A). Again, Morgans emphasis follows precedence. As with other early stages, the observation of

similar larvae between dissimilar adults had beenused as proof of relatedness. For example, thetrochophore larva united annelids and mollusks,and the nauplius larva had connected barnacleswith other crustaceans. In A Treatise onComparative Embryology, Balfour (1885) wrote,There is a high chance of the ancestral historybeing preserved in the foetus or the larvae(p 362). Pycnogonids were commonly groupedwith crustaceans based on resemblance of theirpelagic larvae, the protonymphon and nauplius,respectively. Each is arthropodian, smalland roundish, has a median visual apparatus,

and bears three appendages. During Morgansgraduate career, however, Hoek and Dohrn re-tracted their earlier ideas of crustacean affinity,and independently decided that the differencesbetween the protonymphon and nauplius weresignificantly large. This led them to conclude thatthe protonymphon was a separate derivative of atrochophore-like ancestor. To this end Dohrnproposed a hypothetical transformation of thetrochophore larva into the protonymphon. Although Morgan adamantly disagreed with thisimaginative scenario, the hypothesis continued to

receive attention well into the 20th century,perhaps best exemplified by zoologist Joel Hedge-peths (54; p 15) playful stanza:7

The pycnogonid leads a parasitic youth,andnever grows a large abdomen;A precociouspolychaete perhaps,of arthropods a paedo-morphic omen.

After examining the evidence for both thetrochophore and naupliar arguments of related-ness, Morgan (1891) wrote, an answer is exceed-ingly difficult to give (p 33). Excluding a prioriphylogenetic expectations, the protonymphon is

not affiliated clearly with any other larva. Insteadhe emphasized the uniqueness of the protonym-phon. For one, the protonymphon lacks ananus, not a trace of proctodaeum does thepantopod-larva possess ... no reasonable accountcan, I believe, be given to explain how thisposterior opening could have become lost intransition of the trochophore into pantopod-larva(p 33). Pressured to speculate, Morganargued that larvae may be generated de novo,and thus would reveal little about evolutionaryrelationships (Fig. 5B). Here, Morgan followedalmost exactly a discussion from Balfours (1885)

Treatise of Comparative Embryology (p 363)distinguishing two evolutionary forms of larvae.One form represented a modified version of theadult ancestor of the group, and the second formwas introduced as a larva early in the lineage,conserved among members, yet did not representthe ancestor. In a sense, the first form isreminiscent of recapitulationist theories and thesecond resembles Karl Ernst von Baers ideas ofsimilarity among the undifferentiated early stagesbefore specialization of adult morphology. Finally,Morgan (1890) seems to have created a third larval

type, not discussed by Balfour, and he suggests thepycnogonid protonymphon has evolved in thismanner. In the third type larval forms areaffected by the new characters of the adult ... wemay suppose the newly acquired characters of theadult to be thrown back upon the larval formalready present (p 160). Once again, Morgan isskeptical about the constraint that is assumedto exist during organism development. Hedenounces the tradition of assigning primaryimportance to particular larval stages. Thenecessity of believing that the young forms, insuch groups as Annelids, Crustacea, more nearlyresemble each other than do the adults seems tome an entirely unwarranted supposition. Morgan(1890) argues that because the death rate is fargreater for pelagic larval forms than adults,selection acts more vigorously, and results inrapid adaptive responses at immature stages inanimals with pelagic larva. Therefore, ... in suchgroups as the ones we are discussing,Annelids,Crustacea, etc.,we ought to expect .... the reverseof what we find in such a group as the highervertebrates; viz. that the young forms diverge

7 Joel Hedgepeths stanza was written in the style of WalterGarstang, whose delightful poems on larval forms were publishedposthumously in 1951.



8/6/2019 Maxmen Amy 2008

10/13

far apart and the adults come nearer together(p 166).

In the chapter entitled Criticism of Dorhn andHoek, his opinion remains unaltered: The proto-nymphon shares no significant similarity witheither the trochophore or nauplius. Larval simi-larity previously had been the basis for compar-ison with annelids and crustaceans, respectively.Thus, after eliminating these possibilities Morganpresents the reader with a final option, pycnogo-nids have evolved independently or from anancestor shared with arachnids. Hesitantly, Mor-gan (1890) places the pycnogonids nearer to thearachnids than other arthropods. The very greatdifferences in the adult structure of the groups[pycnogonids and arachnids] indicated no veryrecent origin, but possibly they came in at a timewhen the Arachnids had the first pair of appen-dages chelate, and these were innervated from the

supraoesophageal ganglia (p 34). The trouble-some protonymphon is discarded completely fromanalysis, After the divergence of the pycnogonidsas a group from the general phylum of thearachnids, the pantopod-larva may have devel-oped (p 34).

In the decade after Morgans dissertation, larvalevolution received a good deal of attention. Although some larvae reveal evolutionary rela-tionships, the environmental pressures of aquaticlife can result in convergent morphologies. As Wilson (1898; p 1) stated, many of the mostinteresting and hotly contested controversies ofmodern embryology have been waged in discus-sions of the possible ancestral significance of larvalforms, such as the trochophore, the Nauplius, andthe ascidian tadpole. At the time, little wasknown about the basis of inheritance and theevolvability of ontogeny to thoroughly address

Fig. 5. (A) Figures from Morgan (1890) of the pycnogonid (5pantopod) larva, hypothesized to descend from the crustaceannauplius or the annelid trochophore larva. (B) Three models of possible mechanisms by which a larval form evolves within anygroup of animals. Morgan presents this diagram (1890) and explains that larva may not be important from a phylogenetic pointof view. These conditions include: a primary larval form (A) in which the larva represents the ancestral condition, and the adult

forms diverge; A condition (B) in which adults each share a similar larva a possessed by the ancestral adult Xof the group; and(C) a condition in which the adults evolve and their corresponding larvae are subsequently changed such that a is changed to ayand az as the adult X changes into Y and Z.

A. MAXMEN212


8/6/2019 Maxmen Amy 2008

11/13

these questions. Among others, Morgan (27)continued to think about that subject and wouldlatter use genetic crosses to study the inheritanceof larval characters.

CONCLUSION

In the end, Morgans (1891) conclusion remainsdefeatist. He is skeptical about his own phyloge-netic conclusions, Whether future work supportsor disproves such a hypothesis [pycnogonid andarachnid affinity], it is hoped that it may be useful,if only as furnishing another point of view forlooking at the phylogeny of the Pycnogonids, ormay lead to a more complete study of theembryology of the group (p 35). During the sameMarine Biological Laboratory seminar at whichMorgan (1890) presented his findings, E. B. Wilson

ended his relatively inconclusive discussion ofannelid segmentation stating that ... the presentneed is for new facts, not for new theories. Whenthe facts are forthcoming, the theories will takecare of themselves (p 78).

In the detailed manner characteristic of Morganthroughout his career, he identified the keyissues pertaining to the role of inheritance andvariation in ontogenetic processes, exemplifiedin his discussions of the properties of conservationduring embryonic cleavage and germ layer forma-tion, varied development of similar appendages,

and the evolvability of larval stages. Tofulfill the objective of his thesis topic, he wasforced to overlook weaknesses in the currentsystem. Had he explicitly attacked recapitulationor speculative biology in the thesis, he wouldhave risked inciting influential people as agraduate student and perhaps his burgeoningscientific career would have been prematurelyobstructed.

After graduation, Morgan abandoned the prac-tice of using comparative morphology to uncoverphylogeny, joined the experimental movement,and sought out new methods of obtaining datathat would directly test hypotheses about varia-tion during ontogeny. A transition from descrip-tive/comparative to experimental/mechanisticapproaches was taking hold in the youngerscientific community in the decade after Morgansgraduate research. Morgan himself traveled sev-eral times to the Stazione Zoologica at Naples(1890, 18945, 1900) and allied himself with thosestudying embryological problems for their ownintrinsic merit rather than as a means forphylogenetic reconstruction.

Almost 40 years after his dissertation, Morganwould again emphasize the methodological pro-blems inherent in applying embryological data todraw phylogenetic conclusions. It goes withoutsaying that little or nothing has been contributed

by such a procedure [embryology] to the study ofheredity partly because the problems involved aretoo complex (Morgan, 27; p 594). Just as he laterargued that heredity had to be separated fromembryology for science to progress,8 in his grad-uate research he emphasized that embryology hadto be first disengaged from phylogeny before itcould be understood in its own right.

At several times, Morgan (19) reflectedback critically on the phylogenetic tradition:[The biologist] gets confused and thinks thathe is explaining evolution when he is onlydescribing it and he referred to morphologists

as philosophical anatomists9 and historians(Morgan, 09). Needless to say, Morgan neverreturned to his graduate work on the pycnogonids,nor to phylogenetic studies. He did, however,continue to work on embryology each summerand for most of his final years.10 He never grewtired of observing embryos develop, only ofpresumptuous and speculative answers. Heemphasized continuously that only throughcareful observations of the diversity of forms newavenues for research would emerge. For himthe big questions, such as the role of ontogeny in

phylogeny, had to be broken down into operationalunits and approached as components of a largerproblem. Following this prescription, Morganspursuits would eventually return to problemsof evolutionary theory. On the way, he andhis associates laid the foundations for genetics,developing an operative approach to the study ofinheritance and variation (Sturtevant, 59).

8 For an analysis of the role of embryology in the development ofgenetics refer to Gilbert (78, 87) .

9 The morphologists, or philosophical anatomists, form thesecond great group of students whose activity is a direct outgrowthof Darwinism. The determination of the relationships of the great

classes of animals on the principle of descent has occupied much oftheir time. Two other important fields of labor have also fallen totheir share. The study of development or embryology has beenalmost exclusively pursued by morphologists inspired in large partby the theory of recapitulation. Systematists and morphologistsalike have been evolutionists, but it is a curious fact of zoologicalhistory that until very recently there has been no body of studentswhose interests have been directed primarily toward the problemsof evolution ... Is it not strange, therefore, with all the real interestin the theory of evolution, that so few of the immediate followers ofDarwin devoted themselves exclusively to a study of that process?(Morgan, 09; p 372).

10 In his later years, Morgan developed a marine laboratory atthe California Institute of Technology. In the outline of his plansfor the department, Morgan (27) wrote, Life began in the sea andthere lie some of the greatest historical problems of biology (ascited in Allen, 78; p 335).



8/6/2019 Maxmen Amy 2008

12/13

8/6/2019 Maxmen Amy 2008

13/13

Fahrenbach WH. 1994. Microscopic anatomy of Pycnogonida:I. Cuticle, epidermis, and muscle. J Morphol 222:3348.

Gegenbaur C. 1878. Elements of comparative anatomy. F. Jeffrey Bell BA, translator. London: MacMillan and Co.645p.

Gilbert SF. 1978. The embryological origin of the gene theory.

J Hist Biol 11:307351.Gilbert SF. 1987. In friendly disagreement: Wilson, Morgan,and the embryological origins of the gene theory. Am Zool27:797806.

Giribet G, Edgecombe GD, Wheeler W. 2001. Arthropodphylogeny based on eight molecular loci and morphology.Nature 413:157161.

Hanken J, Gross JB. 2005. Evolution of cranial developmentand the role of neural crest: insights from amphibians.J Anat 207:437446.

Hedgepeth JW. 1948. The pycnogonida of the Western NorthAtlantic and the Caribbean. Proc US Nat Mus 98:157342.

Hedgepeth JW. 1954. On the phylogeny of the Pycnogonida.Acta Zoologica XXXV.

Hoek PPC. 1881. Report on the Pycnogonida, dredged by

H.M.S. Challenger during the years 187376. Voyage HMSChallenger Zool 1167. Jager M, Murienne J, Clabaut C, Deutsch J, Guyader HL,

Manuel M. 2006. Homology of arthropod anterior appen-dages revealed by Hox gene expression in a sea spider.Nature 441:506508.

Maienschein J. 1991a. The Origins ofEntwicklungsmechanik . In:Gilbert SF, editor. A conceptual history of modern embryology.Baltimore: The Johns Hopkins University Press. p 4359.

Maienschein J. 1991b. Transforming traditions in Americanbiology 18801915. Baltimore and London: The JohnsHopkins University Press. 336p.

Maxmen A, Browne WE, Martindale MQ, Giribet G. 2005.Neuroanatomy of sea spiders implies an appendicular originof the protocerebral segment. Nature 437:11441148.

Morgan TH. 1889. A preliminary note on the embryology of thepycnogonids. Johns Hopkins University Circulars 9:5960.

Morgan TH. 1890. The relationships of the sea-spiders.Biological lectures delivered at the Marine BiologicalLaboratory of Woods Hole in the Summer Session of1890. Boston: Ginn & Company. p 142167.

Morgan TH. 1891. A contribution to the embryology andphylogeny of the pycnogonids. Studies Biol Lab JohnHopkins Univ 5:176.

Morgan TH. 1898. Developmental mechanics. Science7:156158.

Morgan TH. 1905. The origin of species through selectioncontrasted with their origin through the appearance ofdefinite variations. Popular Sci Monthly 67:5465.

Morgan TH. 1907. Experimental zoology. New York: The

MacMillan Company. 454p.Morgan TH. 1909. For Darwin. Darwins Influence onZoology Lecture at Columbia University.

Morgan TH. 1919. A critique of the theory of evolution.Princeton University Press.

Morgan TH. 1927. Mendelian inheritance of embryonic andlarval characters. Experimental embryology. New York:Columbia University Press. p 594607.

Neilsen C. 2001. Animal evolution: interrelationships of theliving phyla. London: Oxford University Press.

Oppenheimer JM. 1967. Embryological concepts in thetwentieth century. Essays in the history of embryologyand biology. Cambridge: The Massachusetts Institute ofTechnology. p 162.

Regier JC, Shultz JW, Kambic RE. 2005. Pancrustacean

phylogeny: hexapods are terrestrial crustaceans and max-illopods are not monophyletic. Proc Royal Soc London B272:395401.

Sturtevant AH. 1959. Thomas Hunt Morgan: biographicalmemoirs. National Acad Sci 33:295.

Wheeler WC, Hayashi CY. 1998. The phylogeny of the extantchelicerate orders. Cladistics 14:173192.

Wilson EB. 1878. A synopsis of the Pycnogonida of NewEngland. Tran Conn Acad Arts Sci 5:126.

Wilson EB. 1881. Reports on the results of dredging, underthe supervision of Alexander Agassiz, along the East Coastof the United States, during the Summer of 1880, by the USCoast Survey Steamer Blake, Commander J.R. Bartlett,USN. Commanding: Report on the Pycnogonida. BullMuseum Comp Zool VIII:239261.

Wilson EB. 1898. Cell-lineage and ancestral reminiscence.Biological lectures: the marine biological laboratoryof Woods Hole; In the Summer Session of 1897 and 1898:142.

Zrzavp J, Hypsa V, Vlaskova M. 1998. Arthopod phylogeny:taxonomic congruence, total evidence and conditionalcombination approaches to morphological and moleculardata sets. In: Fortey RA, Thomas RH, editors. Arthropodrelationships. London: Chapman & Hall. p 97107.



maxmen amy 2008

Documents