ZOOLOGISCHE STATION 2U NEAPEL FAUNA and FLORA of the BAY of NAPLES Monograph No. 35 ADOLF NAEF CEPHALOPODA Part I, [ Vol. 1, ] Fascicle | TRANSLATED FROM GERMAN Published for the Smithsonian Institution and the National Science Foundation, Washington, D.C. by the Israel Program for Scientific Translations ZOOLOGISCHE STATION ZU NEAPEL Zoological Station, Naples Fauna and Flora of the Bay of Naples (Fauna und Flora des Golfes von Neapel und der angrenzenden Meeres- Abschnitte) Monograph No. 35 CEPHALOPODA by Adolf Naef Parniwwi( Voll) Pasciele with 62 illustrations and 19 tables Berlin Verlag von R. Friedlander and Sohn 1921/1923 Translated from German Israel Program for Scientific Translations Jerusalem 1972 TT 68-50343/1 This is a Limited Edition by Special Agreement between THE SMITHSONIAN INSTITUTION and the STAZIONE ZOOLOGICA DI NAPOLI Israel Program for Scientific Translations Ltd. IPST Cat. No. 5110/1 Translated by A. Mercado Edited by Prof. O. Theodor Printed in Jerusalem by Keter Press Binding: Wiener Bindery Ltd., Jerusalem Contents Fascicle I German English ee y eke FOREWORD V 1 INTRODUCTION MPG ETICHAIECOMSIGERATIONSY | yceveieieis wee woo ow eiele? okebaton aia cieieis) Seheereters 1 8 2. Methods of systematic morphology ..........+2.eees-sceeeee 5 11 3. The form of systematic- morphological presentation .......... 41 40 4. Preliminary systematic- morphological orientation ............ 44 42 5. The recent genus Nautilus and the fossil Tetrabranchiata ... 55 53 MAIN SECTION GHAP TER} 1. Class Cephalopoda Cuvienays a 44.9.5 6 04s afoy)e ne ieleye 77 72 GHAPTER) 2. Subclass Dibranchiata Owen \, «'-e:e/s:s eo is)el «elawvele 90 84 GHAPTER RS. Order Decapoda Veachy yy. y-).:2 7% eicy-j-t i akenee eel akoqens 109 102 CHAPTER 4. Suborder Teuthoidea Naet \. -iy..i4 oy -styeva cages olen totens 135 125 GHAPMERG 3.. Series Metateuthoidea Naeiis iis’. ssc elas ya cktete 149 137 Appendix: Metateuthoidea Myopsida (d'Orbigny) Naef 165 151 CHAPTER 6. Family Loliginidae Steenstrup ................4- 168 153 @GHAPTER= 7. Genus Loligo Lamarck. .....4 240i ele eie cole stoeicie 193 7) CHAPTER 8. Genus Alloteuthis, (Nach ahs mesiieeeme sein 207 193 CHAPTER 9. Metateuthoidea Oegopsida (d'Orb.) Naef .......... 224 210 CHAPTER 0. Eamily, Gonatidae (Hoyle); Pfeffer; jit: cits a. ire 242 226 Genus,Gomatuis) Gray 1... faite otekervedaereheney 244 228 GHAPTER AL. Family, Bathyteuthidae Pfeffer. fy. (0-6 44's). pe-iel- 201 234 Genus’ Gtenoptery x, Appelloiy ayasetcr ret ary 1: 202 235 GHAPTERA2. Family Enoploteuthidae (Pieffer)). pj.) seca seit) fi Jie 261 244 CHAPTERS, “Genus Pyroteuthis Hoyle ip. ceteyaperlsioter« cafexeres 269 201 CHAP TERM. Genus A bral ica) Grayy iaycyerene stone eyoreystenspenenene nel ofa) «ferre= 279 260 GHAP TERelS. Genus Abra lop sis JOubiniis space) stefeyernetels e}oieyeyerey* 285 266 CHAP TERN G6. Genus Mhelidi ote uw this (tether mate kent. )cy. tal.note 296 Dalal CHAPTER 17. Family Onychoteuthidae Gray ......-.seeeeeeaees 801 282 ill Fascicle II German English page page CHAPTER 18. Genus Onychoteuthis Lichtenstein ........... 313 293 CHAPTER 19. Genus Chaunoteuthis Appell6f ............. 322 302 CHAPTER 20. Genus Ancistroteuthis Gray....-.-.ssseeeeee 326 306 CHAPTER 21. Family Octopodoteuthidae Berry ..........+eeee- 334 313 Genus,O'ctopodoteuthis’ Ruppell .......6.02.% 334 313 CHAPTER 22. Family Histioteuthidae Verrill ..............+.+. 343 321 CHAPTER 23. Genus Callitewpois Verill, s.2 20 teeee coe 352 330 CHAPTER 24,. Genus Histioteuthis d'Orbigny sess vesses5 ese 362 340 GHAPTER 25. Family Brachioteuthidae Pfeffer ......4. 200.0900 370 348 GemliseBrachOoteWthis Verrill 52 sa npeces a see's 370 348 CHAPTER 26. Family Chiroteuthidae Gray ..........sesseeeees 376 -°3054 CHAPTER 27, Genus Chirotewthis, d*Orbigny isi sh.. senses 381 359 CHAPTER 28. Family. Cranehiidae Gray -awiass cece viscsant sie 392 371 GHAPTER 29. ‘Genus, Galitenuthis Joubin= (220 .ca steno ae 398 377 CHAPTER 30% ‘(Genus Liiocratire biaPietier ia ee ine elnrsts «5% 402 382 GHAPTER 3. “Gens, Lea cht a Lesveur 52 aialretiss apts ee ome 406 386 CHAPTER 32. Family Ommatostrephidae Gill ..........+eeeee, 411 391 CHAPTER 33. Genus Illex Steentrup ...cccreccectsocsececcee 429 408 CHAPTER 34. (Genus. Tod aropeis"Glrard | sists «Seisiee eee ace 438 418 CHAPTER. do. “Genus Oimmatoestre pies -d' Orbisny 2 .a.. se. e 445 424 CHAPTER 36, Gents Stren oteminis Verified ates sole 459 434 CHAPTER 37. Family Thysanoteuthidae: Keferstein’ i). oce ses « 463 444 Genus Thysanoteuthis Troscheltecsse%2e...84% 463 +44 CHAPTER 38. Suborder Seploidea Nach «ks scsiseles = ses de'e 682 473 454 CHAPTER 39.. Family’Spinulidaé d"Orbigny '” eitiat ied Siar s-eukelaie 505 485 Genus (ol plea Wain. 9 Nene octave sctanens Giaide earners 505 485 CHAPTER 40. Painily Seplidae Kererstelm. 52% aise «reo obse'e aie oo 519 498 CHAPTER I. “Genusicre pia ilies. tac ipieieum ae a'e inte cis a ae sios: ciateieks 544 522 GHAPTER 42. Family Sepiolidae Keferstem:: 6.105 6.4¢caeG id's bailed 564 543 GHAPTER 43. (Gens ROSS4 ay OWER” cysts da ceas wea wieles aadle 589 067 CHAPEER 44. -Genus Heteroreuth is Gray i... indi aeaiee ages 595 573 GHAPTER 45, Genus Seplo ba (Llédch) ) 2c4..2 bed oie ees 601 079 CHAPTER 46. Genus Rondeletiola* Nadel: (Sloe aes 629 608 GHAPTER Al. Genus Sepletta Naef os. oyencaene cop san tenses 640 619 CHAPTER 48. Order Octopoda Leach Viva tiae ee siete evs a tek etal thoes 655 633 CHAPTER 20, suborder Rolypodoldea Naen a5 sist dneurstee tacos wee 675 651 CHAPTER ‘a0. “Ramily Octopodidae d-Orbi ony ooo ts Aiaysls ae eee oe 681 657 CHAPTER ol: “Gens Oc top ts ting forms may have developed by the simplest meta- morphoses. Thus, the actual form may be derived from the type, which can be reconstructed from the concrete form. The metamorphoses of the type are called ''variations,'' a term which expresses a strictly ideal rela- tionship of the actual to the ideal form. So defined, the ideal form resembles the divine creative idea'' of Agassiz or the theme of a set of musical variations. The type has nothing in common with the idea of an ancestor or ancestral form. It is not a hypothesis nor does it exist innature. Its actual existence is not postulated and is infact an auxiliary concept, a methodical tool. Nevertheless, it should not be an arbitrary, unnatural or fantastic construc- tion, in which case it would be of no value to the naturalist, to whom the nature of organisms has a very real meaning. The type must therefore have a "natural possibility.''"** Even if abstract, the type must be naturally meaningful and closely analogous to observed organisms. It is not a paper scheme without life and color. These difficulties are avoided by deriving the type from the most similar and closely related structure, that is, the type should be constructed within a species, genus or family. Such an approach permits a detailed analysis of all compared facts, considering relationships and adaptations, so that the result will be a lively, clear picture. The ideal type represents a complete, naturally possible, if only abstract, organism which is the most character- istic and dominant of objects studied. As the type resembles the single forms in as many characters as possible (the metamorphosis must be simple), it is represented by the same concept which has the same definition (diagnosis) as the single forms. The type of all snails is a snail and that of all vertebrates a vertebrate. In the circle of related forms, however, the type occupies a prominent or central position from which the individual forms are derived and morphologically interpreted. We name this the mor- phological primacy of the type. The term ''atypical'' refers to what is in contrast to the type. * It has been since observed that individual development does not involve the creation of a desired form on rational grounds. Even superficial observation shows a connection of widely different morphological relationships by actual transformation which reveals the true relationships or bridges the contrast between them by transitions. ** We require this also of transitional stages, i.e., all auxiliary concepts which assist the understanding of organic formation (Naef, 1917, p. 21). 12 The type cannot be a static, immutable or unique form if it is to approxi- mate living nature. Each organism develops, and the typical organization also has its development. Among the higher animals (which will be dis- cussed later) the type invariably is an ontogenesis which can be demon- strated by a sequence of stages. Only in this form can it conform to the strictly practical requirements demanded by systematics. All basic mor- phology is therefore embryology. The concept of homology should also be applied to the natural parts of ontogenesis, insofar as they are stages. In a comparison we speak of homologous stages, adding the time correlation of spatially defined conditions (p.8) to the concept of typical similarity. Homologous are those conditions or stages of typically similar ontogeneses which occupy corresponding places in the course of variation of form (see also p. 20). In the final analysis, however, typical similarity does not apply to the ontogenesis of individual organisms but of entire species, in which the variety of forms is often represented by a multitude of organisms of typically different structure (polymorphism). Higher animals, for example, have two sexes, which together constitute the species. Comparison of such species yields a sexually differentiated type representing stages, as in real species, first the common characters of all individuals and then the specific characters of each sex. Thus, the type concept is generalized sufficiently for this work, the immediate aim of which is the systematics of the dibranchiate Cephalopoda. c. NATURAL SYSTEMATICS OF THE SPECIES The concept of type is the methodical basis not only of systematic morphology in general, but of the systematics of species. This term is generally interpreted in relation to species, but it may have also the wider meaning of an abstract order of phenomena (p. 7). We must define here the concept of species, which is the object of this systematic discussion. Many confusing statements have been made about the nature of the species, although the matter is quite clear. Species are natural communities of reproduction of known organisms. Ancestors and descendants, blood relatives, fertile males and females belong to the same species. Nothing else. This relationship may be substantiated in practice by morphological similarities or ruled out by morphologic differences. This, however, has nothing to do with the concept itself, since there is no morphological concept of species. To distinguish between different species is to consider them as separate communities of reproduction; if we unite them into groups we deny this separation, regardless of the special reasons. We shall not discuss here the difficulties of applying the above rule (e.g., restricted fertility of different varieties, etc.). The systematics of species is based on a study of their typical simi- larities, which results in the formation of a "group" or "systematic category’ composed of species for which a common type can be established. It becomes evident that the typical similarity or relationship of forms is 13 graded in the sense that smaller groups can be united into larger groups and these into still larger units. This follows the same basic principle in that the special types are derived from more general types. The graded variety of forms within a larger group of organisms can thus be reduced to the relationship between their respective types. This relationship can be easily expressed graphically by the genealogical tree, the most perfect picture of natural-systematic relations. The potentialities of the genealogical tree are shown in the following example. FIGURE 2. Ideal genealogical tree with an in- dication of the assumed types. Figure 2 represents the genealogical tree of a number of species. There are 11 species (a—l), each capable of producing a number of special cases. b and c belong to the common type m; a,m,d and e belong to n; n and f belong to 0; q belongs to p. All the species are special cases of their respective types. p is the prototype of the whole group. Starting from p, we reach c via o, n and m, i.e., by a series containing 5 types (p, 0, n, m, c). I use the term ''systematic sequence’ for such a combination (p—o—n— —m-—c). The stages of the systematic sequence cannot be omitted or overlooked in a morphological study. For example, the morphological relationship between c and p can only be established by m,n and o, that between c and h only via p,q, r and s in addition to the above stages. The only way to achieve this is if the typical relationships are thoroughly worked out, i. e., if the genealogical tree is the result of a comparative morphological study. There is no networklike relationship as stated by J. Hermann (1783), and no convergence of sequences of stages into a single type. The systematic sequences proceed from "lower" to "higher'' forms, from the general to the particular, but also at the same time from higher to lower units of the system. We speak, therefore, of preceding and of following stages. In addition to and based on the valid range of the type in the genea- logical tree, these stages express Sequences within which one can distinguish between superordinate and subordinate arrangements. The systematic sequence is, then, a particular case of a transitional series (p.9) and illustrates a metamorphosis of a special meaning already to the idealistic morphologist, not only to the phylogeneticist. (A mere 14 transitional sequence may also be the sequence a—b--e—d—e—f—..)) This is so because the systematic sequence shows the major morphological rela- tionships by which a given structure can be traced back to remote, gen- eralized ancestral types of organic form. Naturally, the completeness of these transitional series depends directly on the existing and available variety of forms. If species g is absent in the genealogical tree in Figure 2 stage r in the series p—q—r—s—h cannot be determined, because this stage results from a comparison between the types g and s (p.10). It might be assumed that the variety of organic forms could be classified also without the apparatus of typical relationships. This is true to some extent: the systematist often uses diagnoses instead of types. A correctly formulated diagnosis is a simple definition which expresses the concept, i. e., the common characters of a systematic unit. A definition contains the higher category (''genus'') together with the specific characters (''differentiae specificae''), providing in each case a complete list of the characters common to all members of the group. Some of these characters are contained in the higher category, others are the specific characters. Nothing else belongs in the diagnosis. The diagnoses form a collective, abstract structure in which each species has its place. They correspond closely to reality, since the similarity of characters is apparently more or less real, and identification does not require any special logical apparatus. Diagnoses can express graded morphological relationships and create a system of organisms which agrees largely with that based on typical relationships. Diagnoses are simpler and would therefore be preferable if they could perform the same task just as efficiently. This is not the case, however. The transition from lower to higher units of the system often involves a rapid loss of common characters. Asa result, the highest units may show no common characters, although the typical relationships remain and permit a morphological characterization to the highest units. No useful diagnosis can be given for all molluscs, nor for some of the subgroups like the Gastropoda, because no common characters exist (Naef, 1919, p.23). The justification of these categories is based on the idea of the ancestral or ideal form (see Introduction, p.4). Although rather vague, this idea can be constructed methodically to a certain point.* On the other hand, there are not natural, i.e., incorrectly established, groups which can be defined by a well-worded diagnosis. An example are the ''Myopsida" of d'Orbigny as will be shown in Chapter V. This group has its common characteristics, in contrast to d'Orbigny's ''Oegopsida.'' The diagnostic characters are, however, typical for all decapods (including the Oegopsida) and are therefore not applicable as characters for the group. Furthermore, the Myopsida are divided into 2 subgroups (Naef, 1913, p. 461), one of which (Loliginidae) agrees with the Oegopsida in some special characters (gladius), and together with the Oegopsida stands in contrast to the type of the other (Sepia-like) subgroup. To distinguish or associate species on the basis of common characters is often unnatural. Such an approach may be preferable for a catalogue of species, but not as the basic principle of scientific systematics. * K. Hescheler (1901) states with respect to the Gastropoda that the genus Pleurotomaria answers all requirements of the ancestral form, so that the systematic homogeneity of the group in terms of our principle is demonstrated. At any rate, the type of Pleurotomaria probably does not include the symmetrical fossil (Paleozoic) Bellerophontidae (cf. Naef, 1911). 15 The artificial nature of diagnostic distinctions and associations is also evident from the fact that they depend entirely on the state of knowledge about the species at the time they are formulated. It is therefore not certain that a given formulation is definitive. Discovery of a new species may invalidate the existing diagnosis or result in quite unnatural distinc- tions. Mammals without teeth, fish without scales, snails without shell and stags without antlers have been found, and the diagnoses would have to be changed to accommodate these ''outsiders.'’ Despite their striking deviations, these forms are classified on the basis of a principle different from any similarity of characters, which shows that this principle has been tacitly if not openly accepted. The obvious disadvantages of the system of diagnoses cannot be corrected by dividing the characters into ''essential'’ and ''nonessential.'' Such a distinction is arbitrary and cannot be used methodically. The diagnosis has not the same function as the type in another sense as well. The type does not only provide the principle of unification of a group of forms, but determines, by centralization of the morphological comparison, an order of ranks by regulating the sum of morphological relationships (Principle of Systematic Procedure, p.19). In the diagnosis everything is equal. As stated by Goethe, the type is learned from nature, and is natural despite its idealistic raiment. The methodically derived formulation is obligatory, based on knowledge which will not be upset by the first puzzling fact. If anew species does not fit into the ideal framework of a family, it obviously belongs to another, possibly wider circle of forms which does not affect the original content and composition of the family.* Conventional systematics with its hierarchy of classes, orders, families, ete. should express the typical relationships of the genealogical tree. This is generally accepted, irrespective of whether the tree is ideal or shows actual descent. The translation of these relationships into conventional terms has resulted in many difficulties which have practical importance if the relationships have to be formulated and systematized. The following is an example. PP PP Species eee Genera W eehte on fe Families OM ANNE n Aone ere any epee Orders bf ------- 20 ----- +--+ ------ Classes of SAAS NS Nain aictelsionie eines Phylum FIGURE 3. Ideal genealogical tree and its transfer to the conven- tional systematics. * The absolute determination of the rank of groups is not natural but fulfills only a practical need. Con- sequently, it can be changed to fit new species. Such a change does not detract from the value of the type or alter the composition of groups. 19 16 Ea Figure 3 may represent a larger group of animals. Translated into systematic terms, it easily leads to phylum, classes, orders, families, genera and species. Real genealogical trees, however, are often quite different: 1. The number of branching points (types) is often greater or smaller than the number desired or needed to obtain systematic clarity. In the first case this is solved by the use of ''intercalated'' units such as subgenera, subfamilies, suborders, etc. These are often insufficient, and it becomes necessary to add tribes, series, and other units without definite rank. If the number of branching points is too small, ''empty' groups have to be used, e.g. families with a single genus, genera with only one species, etc. Such units only fulfill the need for symmetry, clarity and abstract opposition but not the inner need for a systematic relationship. Where there are too many branching points it may not be necessary to give a diagnosis for each unit, especially with respect to subgenera, subspecies and other small categories — but not for species! (Naef, 1919, p. 48). 2. A much more serious difficulty arises when the types or closely related forms actually exist and have to be inserted in the system. To the phylogenetical systematist these forms are just as inconvenient as they are instructive, since they cannot be placed as species in the conventional framework of the system, at least not without compromise. They would have to be coordinated to the other species and at the same time superordinated. A truly adequate form of a natural system should have a well defined place for the existing types. It should also express the gradation and sequence of morphological relationships. This can only be achieved by a genealogical thee: 3. The above problem is connected with another difficulty. We are dealing not with ancestral forms which do not exist but with closely related forms which share some typical characters with the ancestral forms, incontrast to other forms, which are characterized by acommon deviation which is typical for them. As the following examples show, we are dealing with a central group from the type of which the type of a divergent category has to be derived. It is an unconscious convention in systematics to coordinate the whole central group to that derived from it, provided that the type of the latter is more or less contrasting and well defined. Such an approach, however, would be an encroachment on the principle on which the system is based. It would be necessary at least to characterize the derived groups distinctly. To be consistent, the central group would have to be divided into as many branches as can be coordinated to the derived group. This would impair the clarity of the system and necessitate major changes in the existing system, which should be avoided as far as possible. Thus, the Dibranchiata are placed opposite the ''Tetrabranchiata" which, on close consideration, actually form the central group of the former and are divided into several subgroups. With the exception of the Endo- ceras-like forms, these subgroups have to be assumed to be derived from a common ancestral form, which we assume to resemble Orthoceras. However, the same ancestral form belongs to the Dibranchiata as well, while there is no justification to connect the various Nautiloidea and Ammonoidea on the basis of similarities that are not valid for all Cephalopoda (see Introduction, p.5). Hence, the Dibranchiata would have to be coordinated with the natural, questionable subgroups of the Tetrabranchiata. 20 18 In reality, one often has to deal with a number of coordinated groups arranged in a systematic sequence where each group represents the core group of another (i.e. the ''following'') group (p.13). If this relationship has to be preserved, it must be specifically formulated by a) a clear state- ment to this effect and b) a definite sequence. The natural orders of Gastropoda may serve as an example* (Naef, 1911), although the relationships among the classes of vertebrates are better known. Fish, Amphibia, reptiles and birds are not branches of equal rank on the genealogical tree. Actually, in the above sense, each class is the central group for the following class. Thus, the type of the birds is derived from that of the reptiles, the latter from that of the Amphibia, and this finally from the type of the fishes. Furthermore, the major groups of fish (Selachii—Dipneu- sti** —Ganoidei—Teleostei) show essentially the same relationship, although their mutual relations are still highly uncertain. Introduction of these relationships in the genealogical tree gives the following results. The umbelliferous pattern (Figure 3), which is directly translatable into systematic terms, is often replaced by a horsetail pattern in which the steplike modification of the original forms is more evident. This is particularly the case with the Gastropoda and Vertebrata (the phenomenon was first observed in the vertebrates). It is a survival of the early concept (of Bonnet?) of the ''animal series,'' which contains a correct view. The system of animals prevailing today is not strictly like a genealogical tree of the variety of forms (Figure 2). It represents a rather steplike order (p.13) in which no attempt is made to separate the various hetero- geneous principles involved. A natural as well as practically useful system must combine the sequence of systematic or phylogenetic stages with the tree pattern in which the types or ancestral forms are divided. All accepted systems have done this without discussing the matter in principle. Principle: The following systematic categories should be coordinated as of equal rank: a) subgroups derived by divergent variation of a type, or b) all species placed before and after two successive stages of the genea- logical tree, provided there is a distinct morphological contrast between these two stages. In the latter case the sequence is important and must be explicitly formulated. If such a series continues, the contrast can always be used to establish and distinguish successive steps. Examples of such contrasts are Tetrapterygia— Tetrapoda, Anamnia— Amniota, Diotocardia— Monotocardia. Whether types or diagnoses are used to introduce order into a variety of forms, a thorough analytical treatment is indispensable. Thus, the onto- geneses are divided into stages in which we distinguish and compare single characters. In complex organisms it is advisable to follow the development of well defined parts through the sequence of stages, the parts in question being considered as units of lower rankt which possess a relative indepen- dence. To arrive at a diagnosis one merely has to find the common charac- ters, and this does not require any special logical equipment. The common * The groups (orders) of Gastropoda (Belleromorpha, Zygobranchia, Trochomorpha, Neritacea, Ctenobranchia, Opisthobranchia) are staged groups like the classes of vertebrates. ** Including the Crossopterygili: t "Morphogeneses.” See Naef,1917,1919. 21 19 characters of a group naturally also apply to the type. To determine the type, however, the different but homologous characters have also to be compared. This analysis involves differentiation, evaluation and considera- tion of the individual elements. It attempts to distinguish the typical from the atypical and to obtain a picture of the type as a whole by determination of the typical in the individual. d. DEFINITION OF THE TYPICAL AND THE MORPHOLOGICAL PRIMACY The widespread practice of relying on ''morphological instinct" in dif- ferentiating between typical and atypical characters should not be applied, as such an approach never resolves controversial problems. It is also quite possible to formulate clearly the criteria for such an evaluation. A true science cannot rely on emotion to solve its basic problems and at the same time maintain a soul-destroying objectivity. Yet an emotional concept does assist the morphologist. There is a definite union of ideas between esthetics and morphology which deserves more thorough study. The type occupies a prominent place within the ''group'' or systematic category which it characterizes. This position may be named the ''morpho- logical primacy'' of the type. The question is whether this primacy applies also to separate characters and complexes of characters which diverge from the type or whether the type merely shows average values. The latter appears to be more probable. A review of the morphological literature shows that there are numerous animal forms which occupy a peculiarly dominant place in their respective groups but which also show a markedly eccentric character (Pleurotomaria among the Gastropoda, Nautilus among the Cephalopoda, Amphioxus among the Chordata, Hatteria among the reptiles, etc.). These examples show that morphological primacy does apply to certain individual forms, although this concept is not clearly defined. Furthermore, the relationship does not cover all characters of these forms but only an important part of them (p. 23, IV). Morphological primacy is based on three major criteria: the ontogenetic, systematic and paleontological precedence. I. The primacy of systematic precedence states that a character typical for an entire group of forms has to be considered also as typical for the subgroups, provided it occurs in them. As noted on p.13, the type of the larger group is defined in the genealogical tree as a systematic stage which precedes the type of the subgroup. The characters of the larger group appear in the subgroup as a morphological primacy with regard to other characters. Thus we arrive without difficulty from the higher type to the individual forms. It follows, however, that the lower type must be defined in the same manner. This principle follows from the nature of graded typical relationships. Just as the content of the wider diagnoses passes (in the form of a higher concept) into the diagnoses of the subgroup, the higher type determines the subordinate type and the preceding stage the following stage. This is an expression of the inherent connections in a system of organic forms based on typical similarity. Ze Il. The primacy of ontogenetic precedence follows from the general experience that the degree of similarity between morphologically related organisms increases at the beginning of development. This statement can naturally be extended to the type. To use J. Muller's expression: juvenile stages of different animals resemble the common type more closely the younger they are. This formulation needs qualification before it can be considered as a rule, because there are apparently many exceptions, especially in larval and embryonic organs. The larvae of Diptera, for example, differ externally much more than the adults. The nature of these exceptions gives the explanation. Larval and embryonic organs are complete structures, not early stages of parts of the adult animal. The rule deals only with these early stages and their relation to the later condition. This is not a question of mere precedence or even absolute age, but only an observation that certain structures are the early stages of others which develop from them later. The morphological stages which occur during ontogenesis always resemble the type more closely than the stages which later develop from them.* In this formulation, there are 20 no exceptions to the rule (I have not found a single exception in the literature or in the very large material I examined). A more detailed discussion of this principle cannot be made here. Numerous examples may be found in the special part. It is seen from the above that embryology is very useful to systematics since it provides a large number of diagnostic characters. It is of particular interest that embryological phenomena take a different course although they start from identical early stages. It is, however, impossible to dis- tinguish directly between these divergent trends, between typical and atypical characters, on the basis of the morphological primacy of the pre- ceding stages. This would require further auxiliary considerations. The question arises whether the relationship to the typical early stage permits an evaluation about the following stages. An objection to the above statement must be answered here. The simi- larity of the early stages of homologous structure to each other is due not to a lack of specific characters but to the existence of morphological elements which are in sharp contrast to those of the later stages. For example, typical juvenile stages of lungfishes (Crossopterygii) have external gills like the larvae of Amphibia, which they resemble in other respects as well. These juvenile organs later disappear in both groups. There are species in which the typical juvenile states persist completely, i.e. only some details being changed, while the position and arrangement of the major components remain the same. In others the general ''plan" changes radically. The primacy of ontogenetic precedence can be applied to the case in which adult characters developed directly from the typical juvenile condition.** Such a case can be defined as a similarity of style between adult and young. Quite different are cases in which the typical juvenile condition persists (''neoteny''; Naef,1919,p.31); further develop- ment and modification of certain parts is here simply inhibited. * Ihave previously attempted to formulate this qualification by the concept of morphogenesis (1913, 1917, 1919). ** In comparative anatomy certain types have been described as embryonic; this term has also been applied to fossil groups (cf. Broili-Zittel, 4th ed., 1915, p.11). 23 21 The similarity between the early states (‘'anlagen'') of some structures has often been considered as proof of their homology, and this is quite correct. The principle may be stated thatif two structures are homologous in the state of ‘anlage’, they must be homo- logous later also. However, this does not mean that the concept of homology can be based on development, as is generally done. This would be a ''petitio principii.'' We are dealing only with the homology of "anlagen", which can be directly demonstrated to prove the homology of the later structures which develop continuously from them. We must now consider the old theory of the 'parallelism between onto- genetic and systematic sequences" (in my terminology) or the "parallelism between embryology and the animal series,'' in the formulation of Meckel and the idealistic morphologists. This is an extension of our principle which states that the preceding stages are closer to the type of the next systematic group or stage thanthe stages which develop from this type, but the farther back one traces the ontogenetic sequence, the larger the number of forms included in this approximation. K.E.von Baer clearly formulated this observation in stating ''that the earliest stage of each organism shares the greatest number of characters with the earliest stages of all other organisms; that ata slightly later stage the organism is structurally similar to a corresponding stage ina smaller group of organisms; that in each successive stage the embryo acquires new characters which distinguish it from an ever increasing number of groups of other embryos which were previously identical with it, thus progressively reducing the group of embryos that still resemble it; in this manner the group of similar forms shrinks to the single species to which the embryo belongs." Translated into the concepts used here, this means that morphogenetic processes pass more or less closely through the systematic stages which lead to the species of the organism because the occurring forms are characteristic or typical for ever smaller groups of organisms. This shows the enormous value of embryology for natural systematics; if the above law is applied in reverse, an assumed systematic relationship can be checked against the individual development. Ontogenesis shows a real progress from lower (usually simpler) to higher (more complex) forms and establishes a natural connection between them. This gives solid support to the view that the variety of organic forms is internally coherent and that a ("natural") system truly represents the natural relationships (p.6). Systematic morphology thereby gains con- solidation and this explains the deep interest in embryological studies, from the beginning. Ill. The primacy of paleontological precedence was not actually estab- lished — or at least not clearly formulated by the idealistic morphology. This would have presented the variety of organic forms in a historical perspective. Yet even a strictly idealistic morphology could not have ignored the fact that geological evidence shows a distinct precedence of the typical and succession of the atypical. Thus, the Tetrabranchiata preceded the Dibranchiata, the Nautiloidea preceded the Ammonoidea, fish the Tetrapoda, etc. The idealistic morphologist was well acquainted with the facts of this progress. 24 22 23 Haeckel* pointed out that the paleontological sequences correspond strikingly to the systematic and ontogenetic sequences (Reihe der Wirbel- klassen, p.17). However, the paleontological sequences are generally less consistent and complete than the systematic sequences (p.13). By contrast, ontogenesis with its complete series of stages demonstrates the metamor- phosis of the organisms as a real process, and shows the natural connection between the different stages. The gaps in the paleontological picture, however, should not detract from its value since they are due mainly to the lack of preserved material. Thus, both ontogenetic and paleontological data strongly support our systematic-morphological concept. Mainly because of them the ''natural system’ appears as a formulation of facts which must and can be explained, as pointed out in the Introduction (p.5). The theory of evolution provides a causal link between the three series of phenomena. It also reduces the three kinds of precedence to a cause and effect rela- tionship.** The systematic- morphological way of thinking then appears as causal thinking in disguise. In the final analysis, the systematic-morphological way of thinking may find here its deeper justification, and one is tempted (like Haeckel) to discard the methodological apparatus of idealistic morphology and to adopt instead an essentially historical viewpoint, which the proponents of ''phylogenetics" have not done. This is indeed my intention. However, I cannot produce at present a valid equivalent to the principles of idealistic morphology. The principles of historical order of the variety of organic forms can only be based on the laws of phylogenetic development, and these have not yet been clearly formulated.t On the other hand, a methodology of idealistic systema- tics is partly available and the rest we could construct by elaboration of existing views. Our particular purpose can thus be achieved in a logical manner. Let us return to the morphological primacy. The question arises whether the series of principles which determine the morphological primacy can be Supplemented by the inclusion of further elements. In my view, principles of equivalent rank cannot be established any more but it is possible to formulate and tacitly apply some auxiliary principles based on the three basic ones. IV. The primacy of typical correlation means that among comparable states of typically similar organisms, the higher rank in the definition of the typical belongs to those which are naturally correlated to others that are already established as typical (Naef, 1919, p. 30). V. The primacy of complete development. If two organisms are typically similar but some of them lack homologous structures or their final stages, the organisms in which these parts are fully developed rank higher in the definition of the typical (cf. Naef, 1919, p.31; also the chapter on the genus Sepiella in the special part, particularly with regard to the absence of luminous organs). * Haeckel considers the threefold parallel of systematic, embryological and paleontological sequences as... “one of the greatest, most important and most wonderful phenomena of living nature." ** The nature of ancestral forms is the "cause," while the nature of the descendants, which depends (at least in part) on that of their ancestors, is the "effect." t The morphological primacy would have to be replaced by the historical (phylogenetic) priority, for which similar criteria should be established (Naef, 1919, p.51). 25 24 VI. The primacy of monomorphous development of homonomous parts. Parts of the same individual are often typically similar (‘homonomous'"’), and we will attempt to construct a type for these parts as well. This does not involve the establishment of an ancestral form, but only an ideal basic form. The abstraction made in connection with a single part, however, will be applied to the whole. For example, having established that the basic form of the limb of a lower crustacean is a biramous leg, one may assume that such an animal could be completely equipped with biramous legs. A comparative study of Crustacea suggests that such an ideal form is the type, and we ask whether this way of thinking might not have a more general justification. This conception seems fruitful and firmly based in the spirit of idealistic morphology. Thus, wherever a structure is seen to be poly- morphous in the same individual, it is possible to visualize an ideal form in which this structure is monomorphous, i.e. expressed in a uniformly typical manner. This ideal form often represents the general type of an enlarged group of forms, that is of a preceding systematic unit, and we are inclined to give this structure morphological precedence also if it is unknown (extinct?). We therefore consider the construction of ideal forms with monomorphous parts as a heuristic tool of systematic morphology. The structure observed can be derived from such ideal forms according to the principle of the division of labor. Applied in reverse, our way of thinking shows the primacy of undifferentiated polymorphism over differentiated polymorphism. Principle: the ipolymorphous structure of, homono mous parts in. an, organism ‘(or in individuals of.:a) group) ican be reduced to a1 monomorphous. structunmeiin which. the typical structure is: replaced .by .an- atypical strujetunes Uniformly constructed homonomous parts are. to be considered as.primary within a:.systematic category (numerous examples may be found in the special part). VII. The primacy of monomorphous structure in individuals of the same species. Not only individuals of different species and parts of the same individual are typically similar, but also individuals of the same species. The degree of such a similarity may be so great that one can speak of morphological identity and not proceed to the construction of a type, which each individual apparently represents by itself. However, this is not always the case. The polymorphism of species is well known. All textbooks stress, for example, the typical similarity between polyp and medusa, or derive the form of the medusa from the type of the polyp, doing so by postulating the existence of a more remote ideal form of the species in which all individuals are uniform. This is in fact the assumption of the morphological primacy of the monomorphous state. Polymorphous species are always considered descendants of monomorphous species. These considerations can also be applied to sexual dimorphism. The dimorphism of sex cells is undoubted typical (general), at least for all the Metazoa. Monomorphism of these cells will have to be placed in the remote past to the most general forms of dioecious reproduction. It is different with respect to the so-called secondary sexual characters. Here we will have to look for a more closely similar monomorphous prototype which is represented by the basic form of the genus or family in some cases, by that of the class or phylum in others. More information can be obtained by a 5110/1 %6 25 special comparison of forms. A particular dimorphous character can often be traced back to a monomorphous character without assuming a monomor- phous structure of the respective type. An example is the hectocotylization of decapod Cephalopoda, which affects different arms. This could occur in the ancestral form but not in its specific structure. Principle: <1f ‘the individuals*of ome species (or the species ‘of a group) show a different*structure, this con- dition has to be traced back to a monomorphous state which, if it exists, should be considered to have morpho- logical primacy in a systematic category. Neither of the above principles means that the simpler state is the starting point of morphological variation (p.10). Not all complex states can be traced back to simpler ones, if our basic concepts are valid. Many complex types of organizations undergo a successive series of simplifica- tions which leave a distinct mark in ontogenesis. Examples are the develop- ment of the larvae of Ascidia, Sacculina, parasitic snails, etc. The metamorphosis of individual parts often follows a descending path. There are blind forms among vertebrates and cephalopods, the eyes of which are primarily typically differentiated. A principle of progress cannot be taken into consideration in the evaluation of typical similarity. ''Lower'' forms (p.13) are not always simple (p. 21). The last 4 principles (pp. 22—24) are subordinate to the 3 main principles (pp. 19—21) and are valid only insofar as they do not contradict these. In some cases the differentiation of homonomous structures or the polymor- phism of individuals is typical for a group but some of its members do not show these differentiations. This means that metamorphosis reverted secondarily to a state which has to be assumed as the starting point. An example is the uniform segmentation of the body of many insect larvae which has to be assumed for Arthropoda in general, but is not applicable to the special type of insects. If none of the corresponding states of typically similar organisms can be considered to have morphological primacy, the mean value has to be considered as typical. This rule is not important for the determination of types, but it permits their visualization, determining body proportions and other numerical values which could not be determined otherwise. e. ATYPICAL SIMILARITY Systematic morphology deals with the variety of organic forms from the point of view of typical similarity, not general similarity. There are also atypical similarities or ''convergences'' which the morphologist has always found difficult to identify. It is characteristic for these similarities that during the arrangement of the similar structural elements into the respective wholes and during consideration of their development one finds that these similar structures are in contradiction to the systematic position of these forms. Atypical similarities must be eliminated in systematic morphology. The two types of similarity are traditionally known in comparative anatomy 27 26 as homology" and ''analogy.''* Atypical are similarities which approximate the later members of different systematic ontogenetic or paleontological sequences and thus reduce the contrast between the preceding, not or less similar members of these sequences. This is convergence of the series, instead of typical divergence. Atypical similarities are common but do not form a coherent system of relationships which requires interpretation in principle. The existence of these similarities has resulted in the erroneous theory of "network rela- tionships" (cf. also Naef, 1919, p. 54). f. IDEALISTIC MORPHOLOGY AND PHYLOGENETICS Darwin stated that it is not a scientific explanation to make an ideal "ylan'' according to which related organisms are constructed. It cannot be denied that the theory of evolution and the phylogenetics based on it are closer in many respects to natural science than the concepts of idealistic morphology. On the other hand, the theory of evolution has developed from the study of ideal relationships and is based on the natural systematics which developed from this study, and it has to be assumed that each advance of idealistic morphology will result in an advance of the theory of evolution. It has not been proved so far that one can proceed logically in phylogenetical theory without recourse to idealistic morphology. If the organic variety of forms is interpreted and classified in accordance with these views (or a part of it, as will be done for the dibranchiate cepha- lopods in the special part), it emerges as a system of idealistically inter- preted relationships in the form of a genealogical tree (p.12) which can be readily translated into phylogenetic terms. Phylogeny has achieved this and nothing else. To justify the basic concept that systematic rela- tionships are in fact the expression of phylogenetic relationships, phylogeny has produced a series of circumstantial proofs by which the original hypo- thesis was consolidated into a theory (pp. 5—6). Accordingly, the terminology of the earlier idealistic morphology was replaced with a new terminology suitable for the new concepts as follows: Previous term New term Relationship of form Blood relationships Metamorphosis Phylogenesis Type Ancestral form Typical states Primary states Lower animals Primitive animals, etc. These changes, however, are not based on an essential revision of the idealistic morphology. Neither the specific results nor the basic concepts of idealistic morphology were subjected to a critical examination, which was especially necessary because of the revolution of concepts. From its beginning, phylogenetic morphology always moved in the dark, since the material reality in which the more subtle considerations of idealistic * But in a very imperfect manner, since the comparative anatomist is not interested in analogy. Further- more, homologous parts are often also analogous. This approach does not go to the crux of the matter. 28 27 morphology resulted, caused the vague and esoteric principles of this science to be neglected or forgotten. Scientists believed to be dealing with facts,and ideas appeared superfluous. It was not realized that the search into the past from the point of view of historical morphology required a new methodological approach, different from that which had been used in the recognition of ideal relationships or the proof of the theory of evolution. It was particularly necessary to obtain historical evidence and to examine thoroughly the nature, value and reli- ability of the available facts. This has still not been done, and the need for a change in this field determined the character of the present work toa large degree. Morphology is a historical science. Like any branch of history it has to prove to what extent the past determines the present. Closer examination shows that it is impossible to understand the present state of organic beings without reference to their earlier states, especially with regard to structural relationships. On the other hand, the inclusion of such ''early states" in the explanation'' provides a valuable insight which can reveal the rules which govern the conversion from one stage to the next. The historical view forgoes (nolens volens) the examination and determination of these rules and is restricted to assuming that a given state is conditioned in principle by the preceding states and to determining the series of former states. Since every state of an organism is a variable system and can be directly associated only with its immediate precursor (and its energetic situation), earlier relationships can be traced only indirectly. It is an established fact that higher organisms have a history dating back to the mature egg which includes a series of increasingly complex stages, each of which is determined by its precursor. This "ontogeny," or individual ''developmental history,'' is well known in a large number of cases. We also know another type of history, the development of the egg cell itself which exerts a direct influence on the course and results of ontogenesis. We know, for example, that the ovum undergoes certain changes after fertilization, that earlier the ovum developed the so-called maturation phenomena while still in the maternal body or after leaving it, and that still earlier the ovum was formed together with others by a number of divisions developing from a primary ovum which in turn developed from a mature ovum. The ovum thus has a history of its own, having developed by a number of cell generations from an identical earlier ovum. This cyclical develop- ment, perpetually reverting to its starting point, may be termed "germline development. We do not share the apparently resonable view that descendants are the product of the parent body. The germ cells are indeed harbored by the parent, but they do not take part in its biological activity. Historically, germ cells develop directly from earlier germ cells. The fact that ordinary cells in plants or lower animals assume (or retain) the function of germ cells and produce a new multicellular individual does not change the above rule in principle, since also in this case these cells were not produced by the parent organism but participated in the formation of the parent organism. The cells form the cell states, but these do not produce their cells. The existence of a sharp contrast between the somatic and the germ cells in higher organisms, especially animals (Weismann's germ plasma theory) 28 only stresses this relationship but does not create it. Since we are dealing 29 29 here with the phylogenetics of higher animals, this contrast is a fact. Animals capable of regenerating an amputated limb repeatedly cannot regenerate a gonad, although the limb has a much more complex structure. This proves that germ cells have a development of their own and become active only during reproduction, like a root which begins to produce a new shoot. The cyclical- rhythmical development of the germ line in higher organisms resembles the development of living cells generally, especially the Protozoa, the most primitive unicellular organisms. In this light the germ cells appear as the basic or primordial form of all organic development and as the precondition of the continued existence of life. In contrast to the cycli- cal-rhythmical development of the germ line, the development of higher animal and plant individuals which is based on it may be named "terminal": from its conception the organism is doomed to die after reaching a certain complex state of organization, generation after generation. During its flourishing, however, the organism continues the development of the germ line and protects it from adverse environmental factors. The germ line (or more generally the development of the cell) is the true bearer of organic history which guarantees the continuation of life from one generation to another, while the higher organism dies after having fulfilled its function. This is the relationship between the endless development of the germ line and the ephemeral existence of the multicellular organism. Phylogeny, which we consider as the history of higher organisms, represents a continuous series of germ lines and not a succession of countless genera- tions of plant or animal organisms, as is generally thought. An interrupted history is no history at all. Phylogeny is a true history and cannot be represented as a line of ancestors, even if this were directly observed. The members of such a line of ancestors are not stages of a process; the evolution they simulate is only imaginary! Idealistic mor- phology was doubly justified in conceiving the connections between related forms as ideal, as these connections remain ideal also for us! Let us return to our analogy. During the brief period of observation, the "underground root" of the germ line produces only identical shoots. As the theory of evolution postulates, however, the root changes with time. This forces us to assume that the shoots which appear continuously from the advancing root change slowly but constantly from one generation to another. The individuals formed today by the same continuous germ line are no longer identical with those produced previously, and future individuals will differ from those existing at present. The historical course of phylo- genesis can be visualized as a lineage, i.e. a series of successive forms. However, the ancestors do not merge directly with each other, but each dies while (and often before) the next develops and therefore cannot show the true nature of history. However, phylogeny can be represented by a true lineage if we consider each individual ancestor as the result of the state of the germ line at the moment the ancestor was formed. The lineage of ancestors thus becomes a function and symbol of phylogeny. The organic forms we deal with thus have a twofold history: an individual history, or ontogenesis, and a tribal history, or phylogeny. With the above reservations (‘), phylogeny can be regarded as the gradual change of the Species along the line of ancestors. 30 30 Phylogenesis is unfortunately often understood as a line of developed indi- viduals and this is an artificial and harmful restriction of the concept. It must not be forgotten that the existing form of a plant or animal is embodied in the whole typical evolution and not in the developed organism alone. This is especially true of higher animals, whose embryonic and larval development often shows a large variety of stages which later disappear. Omission of these stages in a phylogenetical study would result in false conclusions. To avoid such an artificial interpretation of the facts, phylo- genetic development, i.e. the lineage, must be considered as a sequence of progressively changing ontogeneses. This interpretation results in a revision of the ''biogenetic law." If ontogenesis is a repetition of a something, this something can only be the ontogenesis of the ancestors, and if reminiscences of earlier conditions occur in ontogenesis, they should be interpreted and formulated differently from Haeckel, who assumed a direct ''causal connection" between ontogenesis and phylogeny and considered ontogenesis as ‘nothing else'' than a recapitu- lation of phylogeny (cf. Naef, 1917, 1919). These views on the relationship between development of the germ line and that of the individual are of interest also from another aspect. They eliminate the contrast between relationship and homonomy in typical similarity. ‘Consanguineous' individuals, whether living in succession or at the same time, appear as members of a single whole, branches of the same trunk, like the homonomous limbs of a crayfish. Both are really connected by the continuity of cellular development, and the ''metamorphosis" or "variation" of both appears to be causally based on similar principles. This view is only briefly touched on here. I will discuss it later in connec- tion with a general theory of morphogenesis. Another basic concept of phylogenetics, phylogenetic relationship, needs closer examination. Consanguinity is often discussed in phylogenetics, as if its definition were beyond doubt. In fact there is great difference be- tween the relationship among bred individuals, which can be expressed genealogically, and phylogenetic relationship. Both involve a steplike gradation. However, the gradation of the relationship within a family has nothing in common with that of phylogenetics, which is interpreted more simply. Application of exact genetics to phylogenetics may perhaps change this situation by providing, for example, a closer connection between species formation and the actual relationships between forms. At this stage, how- ever, the phylogenetic relationship is interpreted as a tribal or species relationship, while the individual relationship is presumed to exist in each species, without being considered in detail (cf. p. 31). In its substance, the theory of evolution abolishes the concept of species in its most general sense (pp.11—12). Organisms which today belong to different species but developed from a common ancestor are actually a single species. The proponents (like myself) of the monophyletic origin of life believe that in the final analysis all living creatures evolved from a single species. However, at a given time or during a limited period there is always a certain number of communities of reproduction or species, if only the potentialities existing at that time are takenintoaccount. The species concept will be applied further on in this sense. Because of the extremely slow pace of phylogenetic changes, the period does not have to be precisely determined — years, decades or centuries are but brief moments in the 31 31 32 course of phylogeny, but in which species exist. (In the graphic represen- tation of the genealogical tree, the species existing at a given time are represented by a cross section of the branches and the recent state by the end of the branches.) This leads us to the concept of ''phylogenetic series’ (Stammreihe). In tracing the history of a recent species we must assume that its ancestors always belonged to a single ''species,'' to the beginning of life. At any period of history the ''ancestral form" of a recent species is represented by a stage in its phylogenetic (or species) development. Of the infinite number of such stages we shall deal only with those required to illustrate the changes, as in ontogenesis. Such a succession of ancestral forms will be referred to as phylogenetic series.'' This is quite different from the "lineage'' of individual genealogy, although the two are often confused.* Phylogenetics has no connection with the clarification of the individual genealogy which in the case of dioecious reproduction can only be repre- sented by a complex network, but only with the relationship between species (Hertwig, Werden der Organismen, II ed., 1918, p. 236). The only relationship between different species is the presumed or proven existence of a common preceding state in the ancestral form. The criterion for sucha relationship is the common ancestral species, like the common type in idealistic morphology. This gives a complete analogy to the relationships in monoecious reproduction. The genealogical tree thus becomes an adequate expression of phylogenetic relationship, but it cannot fully illustrate the blood relationship in dioecious reproduction. This proves that the criticism of the use of the genealogical tree in morphology on the grounds that it is taken from genealogy, only in which it is justified, is not correct. The genealogical tree is the only adequate method of illustrating an assumed phylogenetic (or systematic) relationship. Any other representa- tion involves numerous compromises and conventions; the same is true for the usual form of the natural system (p. 15). g. PHYLOGENETIC EVIDENCE Phylogenetic research is impossible without historical evidence. There is naturally no direct evidence of phylogenetic relationships, but the natural system, i.e. the sequence of typical relationships, may serve as an indirect expression of such relationships. This is also true of such unquestionably parallel phenomena as individual development and paleontological evidence. We have thus 3 types of evidence for phylogeny: ontogenetic, systematic and paleontological. The past of an organism can therefore be determined from the following facts: a) the position in the natural system, and also the systematic steplike sequence which led to this position (p. 13); b) the series of its ontogenetic stages, or the ontogenetic gradation from which it developed; * The following terms have to be compared and distinguished: “transitional series" (Uebergangsreihe) (pp.9-10); gradation (Stufenreihe) (p.13); lineage (Ahnenreihe) (p.29); phylogenetic series (Stammreihe). 32 33 c) the paleontological gradation, i.e. the geological age and the order of the geological occurrence of the systematic precursors. There are no other sources! The last two series of phenomena have been partly evaluated by the idealistic-orientated systematic morphology. The natural system thus always contains the nucleus of what can be clearly formulated in any particular case. The three sources of information are, however, not applicable only as a historical interpretation of the three main principles of natural systematics (pp. 19—26), but also to individual cases. In fact, ontogenetic and paleontological data contain elements which permit a historical interpretation independently from the natural system. We shall therefore examine these three sources more closely. a) The systematic evidence, based on the natural system and expressed as a genealogical tree, is based on the main principle of the theory of evolution which states that the degree of typical similarity is a measure, or a result, of the degree of phylogenetic relationship. Systematic steplike sequences, represented by series of ideal types, have to be considered as phylogenetic (i.e. tribal) series in which each type is the prototype of a group of species derived from it. Naturally, any gaps or uncertainties in the determination of the typical are reflected in our idea of what is primary phylogenetically. It has to be mentioned that systematics does not give complete phylogenetic series. The number of phylogenetic types to be ascertained, i.e. the methodically accessible types, depends on the existing variety of forms (p.13); hence the limitations of our phylogenetical know- ledge. (If Octopoda did not exist, nothing could be ascertained about a di- branchiate ancestor of Decapoda! ) b) The paleontological evidence is not simply the sum of observations on extinct species. Extinct animals have to be approached differently from existing species because our knowledge of them is incomplete.* This knowledge is often restricted to geological dating. Paleontology only tells phylogeneticists the time at which certain species, genera, families, etc. appeared and vanished (became extinct or changed into other forms), and such information should be treated with care. Only positive data are certain, while negative data, i.e. the absence of certain types in certain strata, are often due to insufficient knowledge of the extinct group. Only a small fraction of the large variety of extinct forms is available, even if organisms that could not be preserved are excluded. Paleontology does not provide direct evidence for a line of descent. Continuous series of transitions have only been observed with certainty in a few cases in close proximity, but the discovery of such phylogenetic series is of more importance as a proof of the theory of evolution than as a specific method for determining the general relationships. Even complete pale- ontological information does not present the individual facts in their organic context (i.e. as lineages). The linkage of these facts into true phylogenetic sequences is generally hypothetical. From the phylogenetic aspect, pale- ontology is thus placed at a disadvantage with respect to ontogenetics, in which morphological relationships can be interpreted historically as natural and coherent connections. * A historical fact is not that something existed but the proof that something existing in the past was the preceding stage of a successor. 33 34 c) The ontogenetic evidence is demonstrated in the individual develop- ment of organisms. Since F. Muller and E. Haeckel, this process has been considered as a shortened recapitulation of phylogenesis. Idealistic morphology proved that the ontogenetic stages correspond to the systematic stages, more exactly to the homologous stages of the preceding systematic stages (p.11). Translation of these systematic stages into phylogenetic stages requires a new formulation of these ontogenetic- phylogenetic relation- ships, as done by Haeckel rather more suggestively than precisely in his biogenetic law. Ontogeny is not a ''repetition' of phylogeny since this (p. 28), whether symbolized as a lineage or genealogy (p.29) or represented in its true form, is basically different from ontogeny. Ontogenesis repeats the preceding ontogenesis of the ancestors which these phenomena resemble. According to phylogeny, however, there is a gradual divergence from one ontogenesis to the next. This divergence is such that the earlier stages of the ontogenesis are increasingly more conservative than the following stages (with some reservation: see p.19). If this is a law, the observed morphological relationships should necessarily follow. The validity of this law is evident from the general causal connec- tion throughout the development. I have attempted a more precise formula- tion of this law elsewhere (Naef, 1917, p.57). We may therefore assume that every stage of individual development of an organic form must be more primitive than the following stage. Thus, the structure of the different parts of a developing individual repeats similar structures which occur in homologous stages of the ances- tors and appear in the same order as these in the lineage. This formulation describes the evidence of the so-called ''biogenetic law'' and expresses more adequately than this the implications of the true law outlined above. Onto- genesis must thus be considered a well established historical document, but it also provides the key for its interpretation. Although ontogenesis gives no direct information about the adult ances- tors, it does aid indirectly. The transitional forms observed in ontogenesis must have been destined to become other final forms, and their fate must have been different from what is observed today. This is the more probable in view of the fact that the same transitional forms are present in 4 large variety of organisms. For example, entire embryos of higher vertebrates (man) strikingly resemble those of lower vertebrates (shark). FIGURE 4. Ontogenesis (morphogenesis) as phylogenetic evidence. 0-4 are the stages of a morphogenesis with an indirect course. Each of these stages permits the construction of typical final states by direct development. This results in a series of transitional forms (I— IU) which gradually leads to IV. If IV is a bird, HI could be a reptile, II an amphibian and I a fish. 34 35 The transitional forms occurring are not equal in their apparent propen- sity to assume their different course of development, which is direct in some cases and indirect in others. Direct development involves complication and functional development of the parts without affecting the general plan, while indirect development involves changes of the primary topography and proportions, as well as a differentiation and division of function between originally equal structural elements. We consider the forms with direct development as the original forms, or in idealistic terms (p. 20) as typical; those with an indirect development are phylogenetically secondarily changed. Thus, indirect ontogenesis gives at least a suggestion about the structure of the ancestors of the given organism (Figure 4). The sequence of stages of morphogenesis indicates a true lineage in which the respective ‘'anlage'’ states developed once directly in the ancestors, while they developed further changes in their descendants. However, ontogenesis is not a direct repetition of final stages of the ances- tors; there is no palingenesis, in the sense of F. Muller and Haeckel, and therefore no cenogenesis either. Individual ontogenesis can be considered as phylogenetic evidence quite independently of the fact that the comparison of ontogeneses is very impor- tant in natural systematics in which it assumes historical significance as well (pp. 19—21). The advantage of this phylogenetic evidence is that onto- genesis clearly shows the steps of a necessary natural connection between the stages, while paleontology and systematics only suggest such a connec- tion, but do not contain it as tangible facts. Correctly interpreted, onto- genesis gives the most reliable information on the special phylogeny of a given form. However, the scope of this information varies from one case to another and is limited in direct development, in which the early stages follow the later stages in their change. One of two closely related forms may still show a phylogenetic reminiscence which is absent in the other (Naef, 1917, p.60). F.Muller (loc. cit., p.7) speaks correctly of an obliteration of the evidence. There is, however, no falsification of the evidence. If we discussed above phylogenetic evidence separately from systematics, this does not mean a genealogical treelike relationship which is not contained in the idealistically orientated natural system. The whole realm of known facts and the relationships between them have already been exhaustively studied by the idealistic methodics, and phylogenetics could have restricted itself to translating and interpreting the results obtained by idealistic morphology. A phylogenetic evaluation of ontogenetic and paleontological data is nevertheless possible without discussion of the systematic- morpho- logical relationships and even beyond them (Naef, 1919, p. 62). Can the graded similarities be interpreted as direct historical evidence, as phylogeneticists have often done? The answer cannot be negative in principle. However, recent knowledge does not permit the use of simi- larities as evidence for blood relationship (pp. 6, 22,26). I have chosen a logically irreproachable way in elaborating first the principles of the well tested idealistic systematic morphology for consistent practical use. Given such a solid foundation (a theory, no longer a mere hypothesis, see p. 32), phylogenetics can be applied to specific systematic- morphological tasks, as in this book, without taking up the controversial question of the course and causation of species formation. The following part describes the uncertainty which still persists in phylogenetics. 35 36 h. GENERAL PRINCIPLES OF PHYLOGENETIC VARIATION 37 Phylogeneticists often have the tendency to consider the whole change of organic life as progress, i.e. as a continuous, only occasionally interrupted rise from a dark abyss to ever greater organic perfection. Darwin's theory of the survival of the fittest in the struggle for survival has advanced this view considerably. There is no doubt that many steplike series show a progressive increase in size, strength, complication and variety. But this is not always true, since the opposite, i.e. progressive reduction and simplification, often takes place. I hope to be able to show that cases of such retrogressive development are much more frequent than progressive cases. This may seem paradoxical since geology shows a continuous progress of organic life.* Progress which affects only a few lines already appears as a striking phenomenon, which may mask all other trends. Thus, the retrogressive characters recede into the background despite their greater frequency. An innovation like the modification of some of the suckers of decapods into hooks, has to occur only once (in an ancestral species) and may have dis- appeared frequently during or after the formation of new species (Chapter 3). It must be borne in mind that our methodics does not permit us to trace the change of species, but only to determine the systematic and phylogenetic stages (pp.13 and 32) at which the respective groups of forms diverged. Progress provides new possibilities which may become starting points for new, even if rare, variation. Many,perhaps all, phylogenetic stages will therefore correspond to forms which are particularly progressive in some respect. Progress in these stages means the appearance of new variety, but not necessarily a change of species. Darwin's principle naturally does not mean survival of the most pro- gressive forms in our sense, but the survival of the best adapted to an existing or new environment, possibly the smallest, most wretched and humblest form. This is an ecological, not a morphological principle. Morphological progress is often parallel to ecological progress. The two courses have multiplication and differentiation in common. Multip- lication means ecological strength and security. It is associated with the cells, the ultimate known units of life. Cell division is a prerequisite for all progress. It results in multiplication of individuals and their parts, and thus provides material in which differentiation can occur. Differentiation means at first only the appearance of a new form. But, as each form is an "apparatus", it also means a new function and if this proves true, a division of labor. All differentiations dealt with in systematic morphology have this ecological character. In this sense, ''progress' can be defined as a frequently occurring course of development which is not unique but has its opposite and consequently cannot serve as a starting point for methodical investigation. i. PRINCIPLES 1. Idealistic morphology was not only the precondition for the introduc- tion of phylogenetics in the history of science, but has to be given priority * Progress is interpreted here only as complication and an increase of the degree of variety. 36 38 over it on logical grounds even today by creating the ''natural system" of organisms (we cannot investigate things that do not exist any more; see Pp. 6225216) - 2. We consider the theory of ''typical similarities" as the methodological foundation of idealistic morphology. Typical similarity is the similarity between complex units which consist of similar parts in the same arrange- ment (pp. 8—11). 3. Typical similarity exists between organisms in general, especially with regard to their forms, functions and adaptations. Systematic mor- phology is the study of organisms in terms of typically similar forms (pp. 7—8). 4. Typical similarity is analogous to geometrical similarity, particularly in the sense that each part of a structure is "homologous" or ''corresponds'' to some part of another structure. Typical similarity is thus an ideal relationship between forms, which is expressed in the homology of their parts. In the final analysis, however, typical similarity exists not between fixed but between developing forms, i.e. between entire ontogeneses to which these forms belong (pp. 8—11). 5. The type is primarily a''tertium comparationis" in the comparison of typically similar single forms, and relates a multitude of such forms to an ideal center. The type appears as an expression of objective necessities in this application. The relationship between individual forms and the type is comparable to that between individual cases and the law (Goethe). 6. The typical is determined by comparison of the individual parts with observance of a ‘morphological primacy" of certain parts or characters. This is done according to certain principles (p. 18). 7. There are 3 main principles of morphological primacy: ontogenetic, paleontological and systematic precedence. All three develop historically- dynamically from the primacy of cause over effect (p. 22), but they can also be derived directly from the morphological relationship of the forms ob- served and thus be based on idealistic morphology (pp. 19-22). 8. The primacy of ontogenetic precedence means that in the course of a morphogenesis, an earlier stage is more important in the determination of the typical than a later stage which developed from it. This is a more pre- cise and restricted formulation of the so-called ''biogenetic law'' and the old theory of the parallel between ontogeny and the "animal series.'' Earlier stages are to be considered as more important only insofar as they are the cause of the following stages (pp.19—21). 9. The primacy of paleontological precedence means that among the corresponding states of typically similar organisms, those which are more important in the determination of the typical are the ones which occur in earlier geological periods. This can be based on the fact which is proved by comparison based on other principles that such forms are closer to the type (p.21). 10. The primacy of systematic precedence means that among the cor- responding states of typically similar organisms, those which are more important in the determination of the typical are the ones which occur in the preceding systematic rank, i.e. those which are typical for this category. This follows from the logical nature of the systematic sequences. These proceed from larger to smaller categories and what is valid in the larger unit must also be applied in the smaller unit (p. 19). 37 39 11. Embryological facts provide the most reliable information on the typical and its relationship to the "derived". There is in ontogeny a gradual transition from certain basic or primary forms to a variety of derived states (p.21). 12. From these considerations emerges the concept of metamorphosis, in which observed (and ideal) individual forms (including adult forms) are arranged theoretically in a series, like the stages of a process. Metamor- phosis is the ideal combination of typically similar forms by transitional stages (transition series) (pp. 9—10). 13. By such a process, i.e. by a series of transitional stages, the indi- vidual forms appear to be connected with their type. The type is then the abstract (but naturally possible) individual form from which a multitude of typically similar forms can be "derived" by the most simple and naturally possible metamorphoses (p. 10). 14. Typical similarity is the foundation of the natural system since it places the forms (species) related to one type in a single systematic cate- gory. If such types are again typically similar to each other, the first formed categories are united in a higher category with a higher type (p.12). 15. The relation between coordinated and subordinated categories is best represented graphically in a ''genealogical tree,'' which does not necessarily reflect phylogenetic relationships. Such a genealogical tree shows most clearly the state of abstract coordination and subordination (systematic sequence; pp.12—13). 16. Systematic categories can usually be characterized by mere defini- tions ("'diagnoses''). However, these definitions have less practical value than the types. A major disadvantage of diagnoses is that they do not show the correlation, i.e. the indisputable connection of characters or parts with the natural whole (pp. 13—15). 17. According to the theory of evolution, typical similarity of species (or relationship of forms) results from their phylogenetic relationship (blood relationship) (p.26). It is therefore justified to transfer the relationships established by idealistic morphology into a historical frame in which the morphological characters of the ideal type are identical with those of a real ancestral form. Phylogenetic morphology can thus be established on the basis of idealistic morphology. 18. Like all true natural science, phylogenetics attempts to explain or reveal the causes of organic morphogenesis. This attempt is, however, restricted to the historical aspect, i.e. the determination of the early states from which a certain form has developed and to which it can be traced back (p. 27). 19. Phylogenetics is thus intimately related to ontogenetics, which describes the directly preceding stages of the given forms and traces them back to the mature egg. Phylogeny can thus be defined as the prehistory of the egg and it culminates in the analysis of the germ line, which ensures continuity of life and form. Phylogeny deals thus with the changes of the hereditary mass during the development of the germ line (pp.27—29). 20. Phylogeny cannot be represented as a lineage or a sequence of ancestors since these ancestors are not stages of a process and are only indirectly related. An evolution which apparently exists in this process is only abstract (in the sense of idealistic morphology, to which we always have to return) (p. 29). 38 40 21. Phylogenetics cannot produce a new concept of homology either. If we relate the various structures to corresponding parts of an ''ancestral'' form, we merely state that these structures are homologous to their ancestral counterpart, i.e. we take the concept of homology for granted. Neither can embryology produce a new concept of homology (pp. 8, 20, 28). 22. Lineages of ancestors can nevertheless symbolize phylogeny, since the structure of the ancestors can be considered the result of the condition at which the germ line had arrived at the moment of its appearance (p. 27). Even in this sense, however, the lineages represent only sequences of developing organisms, i.e. successive ontogeneses. 23. With this interpretation, ontogenesis cannot be a repetition of phylo- geny, as the so-called "biogenetic law'' maintains; this applies also to cenogenesis, etc. There is no palingenesis in the sense of E. Haeckel, in which the adult ancestral stages are repeated in some fashion (pp. 33—34). 24. The structures observed in some parts of the developing individual repeat similar, homologous structures which existed in homologous stages of the ancestors. These structures appear during the transformation of these parts in the same order as they did in the lineage (p. 33). 25. The special significance of paleontology for phylogenetics consists in the fact that paleontology can date fossil forms which have to be treated like the recent forms. The disadvantage of paleontology is that it does not present the natural connection between the facts, which embryology does. All sequences between fossils have to be constructed (pp. 32—33). 26. Phylogenetics abolishes the species concept in its old, generalized sense. Species are only communities of reproduction which exist for a limited period. The majority of such species are "phylogenetically related" (hypothetically) in the sense that they have developed from a common ancestral species (p. 30). 39 4] 42 3. THE FORM OF SYSTEMATIC- MORPHOLOGICAL PRESENTATION We shall not discuss the application of the above principles in systematic- morphological research for various reasons. However, their significance in the presentation of results will be considered. In the preceding part we attempted to explain the abstract meaning of certain formulations and ter- minologies which are widely but inconsistently used. It is evident from the above that an ideal relationship of typical similarity must be the leading factor of a systematic work, while the mere individual facts can provide the material but not the organizing meaning of sucha study. A treatment based on graded morphological relationships, like types in a genealogical tree, is a more coherent whole, superior to a mere collection of data. Such a concept places each fact in its proper place which clearly reflects the degree of its generality or its special characters. Variety thus becomes a unit, and factual knowledge becomes an insight. Such a representation is only possible if there is thorough factual know- ledge as well as completeness of morphological abstraction. Doubts and gaps in our knowledge will persist but a certain completeness within the existing possibilities can and must be achieved. This is a laborious, time- consuming endeavor, ill suited to all the other urgent tasks before me. The presentation must be deductive, proceeding consistently from the general to the particular and following the stages of the natural system to species, and it must show the latter as particular metamorphoses of the generic type. Although this approach is self-explanatory and natural, it has not always been consistently followed, and even in this book we have applied it only to a certain extent. A certain uneveness of presentation could not be avoided. Such deviations will be indicated in the text, for having criticized the lack of principles in others, the author would not like to see his own principles associated with a faulty presentation. The chapters of the following main part describe systematically ideal and concrete species with similar morphology. If the systematic units of each order are based on ideal types (p. 12), they must be described according to these types. Each main chapter therefore contains a description of a species as the type of a genus, family or order. The description of each type is preceded by a diagnosis of the respective systematic category, first, to serve a practical need and second, to give a brief summary of what is found constantly and generally in the whole group. All characters mentioned in the diagnosis are omitted in the discussion. This facilitates the description of the type, since a general character is necessarily also typical. The diagnoses therefore contain important infor- mation on the determination of the type and reduce to some extent the problematic image to which the type is naturally prone. The preceding diagnosis makes many stereotypic considerations unnecessary and introduces a strengthening element in the structure of a systematic- morphological presentation. 40 43 For the sake of clarity, adult forms described in the larger and general chapters are treated separately and before the transitional postembryonic stages. A division into many special chapters would have been chosen if the factual material had been of ideal completeness and had permitted morphological abstraction. We would have then attempted to proceed from the youngest to the oldest stages and would have placed one ontogeny next to another, regardless whether the particular development is typical (ideal), hypothetical (inferred)* or actually exists. However, separate treatment of the adult forms has its advantages. Heterochronisms of development are of less systematic interest. The manner in which typical details emerge is always the same, and morpho- logical variety reaches its peak in the latest stages. Hence the dominating importance of the description of the adult stages. Data on ontogeny only constitute a natural addition to the description of the final state, presenting it as the result of an accomplished process. The description of the typical is followed by a discussion of the 'meta- morphosis" or "variation" of the type in the respective group. This approach stresses the inner coherence of the system by describing not only the unity of form but also its variation. Auxiliary ecological considerations explain metamorphosis as a process of adaptation. The terminology used is mainly that of idealistic morphology, but it can be readily translated into historical terms according to the above principles. Indifferent (ambiguous) terms are also frequently used. An example is "»yrototype'' (Urform), which can mean either ''type'' (ideal form, Typus) or "ancestral form'' (Stammform). Description of typical similarities is always stressed; a strictly historical presentation was used only in cases in which extensive use of paleontological data was attempted or was possible Clarity is a general demand. Pictorial presentation of the observed forms is doubtless worth far more than long descriptions in a systematic morphological work. We used therefore a drawing technique which permits the publication of a large number of drawings at relatively low cost. The specialist will find that the figures form a major, perhaps the most valuable, part of the work. They are drawn after carefully selected individual pre- parations or living specimens as far as their measurements and proportions are concerned, not only reproductions of objects. Inessential and accidental details are omitted. The typical in their outline, symmetrical proportions, etc. have been stressed, often at the expense of the artistic effect. However, also ideal concepts, obtained by morphological abstraction, had to be presented insofar as they referred to relations of form. Their omission would have meant that the ideal concepts have no practical value and have no place in a scientific work and that one should cease to speak of prototypes and ancestral forms and their properties. But if these structures are useful auxiliary tools of methodical morphology and systematics, they deserve ap- propriate treatment. To save space and labor, only types of major groups have been drawn; for the smaller categories the reader is referred to figures in which individual typical characters are illustrated. * The actual existence of which in the past is assumed or maintained. ** On the historical interpretation of the ideal relationships and the “genealogization of the system" see concluding sections 1 and 2 of this volume, as well as my article "Palaeobiologie und Phylogenetik" in Abh. theor. Biol. , Schaxel, Berlin. 1920. 4] 44 4, PRELIMINARY SYSTEMATIC- MORPHOLOGICAL ORIENTATION Contents: a. Definition, position and division of the class (historical). b. Systematic review of the main groups. c. Systematic review of families and genera. d. Scheme of the organization of molluscs. e. Normal orientation of the body. a. DEFINITION, POSITION AND DIVISION OF THE CLASS Aristotle, who had an astonishing knowledge of Cephalopoda, placed them in the class ''Malakia'' and always treated them separately. Linnaeus took a step backward by placing the Cephalopoda partly in the ''Vermes Testacea' and partly in the ''Vermes Mollusca.'' The Cephalopoda were again united by Schneider (1784). Cuvier gave the class its present name, which became generally accepted (1798). However, he placed also the fossil Bellerophon- tidae as well as the Foraminifera in the class. Blainville (1825) recognized the Bellerophontidae as Gastropoda, while Dujardin (1835) transferred the Foraminifera to the Protozoa. The definition of the class has since re- mained unchanged, except for an attempt by Lankster (1885) to include the Pteropoda. Cephalopods are generally considered as the most highly developed branch of molluscs. This is true in terms of morphological complexity and variety; the question is whether there is a special relationship with one or several other classes, and whether the organization of Cephalopoda can be derived from a "lower class.'' Such attempts have been made, of which only two will be mentioned. Grobben (1886) considered Dentalium as a starting point, and Simroth (1905) the snails. Both attempts are based on vague similarities of certain details and are fantastic, without determining the typical general organization of both groups and without methodical comparison. I shall not enter into a detailed discussion of the relationships between the major groups of molluscs; a work on this subject will be published elsewhere. My long experience with molluscs has shown that the recent forms can be divided into 6 clearly defined and equivalent types, none of which canbe derived from another of these groups. These are the Cephalopoda, Gastropoda, Lamellibranchiata, Scaphopoda, Placophora and Solenogastres. 45 The relationships among these 6 classes, to which I add the fossil Odonto- morpha (including Hyolitha, Tentaculita and Conularida) cannot be based on a reconstructed primary type, but only on a more indirect connection because each class shows definite typical similarities with at least one other class. No generally valid diagnosis can be given. 42 46 Critical analysis of the data provides facts which cast doubt on the prejudiced concepts of the primary organization of molluscs. Based on some species of Gastropoda and Chiton, these theories consider the Patella type as the primary form of molluscs. It was shown long ago, however (cf. Naef, 1911), that all Patella-like snails evolved from species with a normal spiral shell and are not related to any primary form of snails. Chiton, on the other hand, is far too problematic in its entire organization, particularly the structure of the shell, which determines the habitus, for any conclusions to be based on it. The accepted theories are therefore without foundation. Some closer relations certainly exist among the six recent classes of molluscs. Solenogastres and Placophora differ sharply from the other classes. This has led to the division into Amphineura and Eumalakia (Naef, 1911), the mutual relationship of which is very problematic. This is a great difficulty for a general orientation of the morphology of molluscs. However, the Cephalopoda form a more closely related systematic unit with the Gastropoda, Lamellibranchiata and Scaphopoda, in which the Cephalopoda and Gastropoda are most closely related. The Cephalopoda provide a con- nection between the above four classes on one hand and Chiton and the Amphineura on the other. The relationship between molluscs and other Metazoa is still very de- batable, but it appears* that molluscs originated from the Annelida. This view is based on basic and indisputable similarities in the embryology of the two groups (see Naef, 1913). The general division of the class has already been determined in general, and I hope to be able to complete it. Aristotle correctly distinguished between the Octopoda and Decapoda, and this was revived by Leach (1817). Owen (1836) established the Tetrabranchiata and Dibranchiata as two opposite groups, and this resulted in the recent classification of the Cephalopoda. Reinhardt and Prosch (1844) introduced the Pteroti and Apteri as suborders of the Octopoda; Orbigny (1845) proposed the suborders Myopsida and Oegopsida for the Decapoda. This division is omitted here as artificial, while the classification of the Octopoda is retained in its content, but in a new and expanded form (Palaeoctopoda, Cirroteuthoidea and Polypodoidea). Other classifications are those of Hoyle (1886: Lioglossa and Trachyglossa) and Grimpe (1917: Cirrata and Incirrata)(see chapter on Octopoda on these groups). The division of the class is based on the historical development outlined above. A new element is the classification of the Decapoda into Belemnoidea (belemnite-like), Teuthoidea (calmar-like) and Sepioidea (sepia-like). This classification will be discussed in greater detail in the main section. The Oegopsida are retained as a subdivision of the Teuthoidea, whereas the '"Myopsida" are abolished. * See the article of K. Heider on invertebrate phylogeny in "Kultur der Gegenwart," Part 3, Section 4, Volume 4. Abstammungslehre, pp.504—511. 43 b. SYSTEMATIC REVIEW OF THE MAIN GROUPS Class CEPHALOPODA Cuvier, 1798 Subclass I. Tetrabranchiata Owen, 1836 Subclass Il. Dibranchiata Owen, 1836 Order A. Decapoda Leach, 1818 Suborder 1. Belemnoidea (Zittel) Naef, 1912 Suborder 2. Teuthoidea Naef, 1916 Suborder 3. Sepioidea Naef, 1916 Order B. Octopoda Leach, 1818 Suborder 1. Palaeoctopoda nov. Suborder 2. Cirroteuthoidea nov. Suborder 3. Polypodoidea nov. ce. SYSTEMATIC REVIEW OF THE FAMILIES AND GENERA OF DIBRANCHIATA Order I. DECAPODA Leach, 1818 Suborder 1. Belemnoideat (Zittel) Naef, 1912 Family 1. Aulacoceratidae nov. Aulacoceras Hauer Asteroconites Teller Dictyonites Mojs. Calliconites Gemm. Atractites Gimb. Family 2. Phragmoteuthidae nov. Phragmoteuthis Mojs. 47 Family 3. Belemnitidae Blainv. Belemnites Lister Ostracoteuthis Zittel Neohibolites Strolley Bayanoteuthis M.-Chalmas. Styracoteuthis Crick Acroteuthis Stolley Oxyteuthis Stolley Macroteuthis Abel, etc. Family 4. Belemnoteuthidae Zitt. Belemnoteuthis Pearce Diploconus Zitt. Conoteuthis Orb.(?) 44 Family 5. Xiphoteuthidae nov. Xiphoteuthis Huxley Family 6. Vasseuriidae nov. Vasseuria M.-Chalmas* Suborder 2. TEUTHOIDEA Naef, 1916 a. Prototeuthoidea nov. t Family 1. Belopeltidae Naet ** Belopeltis Voltz (for 'Loligo' aalensis=bollensis Zieten = "Geoteuthis' bollensis Minst. Family 2. Geoteuthidae Naef Geoteuthis (Belopeltis') simplex (Voltz)=G. lata= G.orbignyana Miunst. Family 3, Leptoteuthidae nov. Leptoteuthis H.v.M. Family 4. Plesioteuthidae Plesioteuthis Wagn., Para- plesioteuthis Naef (for "Geotheuthis' sagittata Munst., see p. 143) b. Mesoteuthoidea nov. t Family 1. Trachyteuthidae nov. Prac ay ULeqGnsS: Ebonvervi, Glyphiteuthis Reuss Family 2. Beloteuthidae nov. Beloteuthis Miunst., Phyllo- teuthis Meek and Hayden (uncertain) Ptiloteuthis Gabb. (uncertain) Family 3. Teuthopsidae nov. Teuthopsis Deslongchamps! (not Teuthopsis Wagner) Family 4. Kelaenidae nov. Kelaeno Miinst. (This peculiar form differs markedly from the typical condi- tion of this group.) The characteristic displacement of the cone to the dorsal * Contrary to the view of Leriche (1906) and Abel (1916), this genus is distinct from Belosepi- ella Alessandri. Belosepiella lacks a phragmocone and has only a Beloptera-like sheath. In V asseuria, on the other hand, the sheath contains an alveole and is distinctly of the belemnoid type. «= The family, originally named Belemnosepiidae, also included Bele mnosepia Buckl., Palaeosepia Theodori, Lolignites Quenst. and Sepialites Miinst. These genera probably belong to Geo- teuthidae Naef. t "Teuthopsis" Wagner is not "Teudopsis" Deslongchamps. This name was originally used for T. agassizi, which was later renamed Loliginites coriaceus Quenst. (Figure 61d); it was later applied to a form closely related to Beloteuthis bollensis (Zieten, Table 37), both from the Lias. For "Teuthopsis" Wagner I propose the name Palaeololigo and the Family Palaeololiginidae (see p. 145). 45 48 ™ surface creates a modification of the mantle sac as in Opisthoteuthis*) c. Metateuthoidea nov. a) Metateuthoidea myopsida (Orb., 1845) nov. Family 1. Loliginidae Orb., 1845 Loligo Lam.,Sepioteuthis Blainv. Alloteuthis nov.,Loliolus Strp. Doryteuthis Naef Family 2. Promachoteuthidae Naef, 1912 Promachoteuthis Hoyle Family 3. Lepidoteuthidae Naef, 1912 Lepidoteuthis Joubin** b) Metateuthoidea oegopsida (Orb., 1845) nov. Family 1. Gonatidae Hoyle, 1886 (as subfamily) Gonatus Gray Family 2. Benthoteuthidae Pfeff., 1900 Benthoteuthis Verrill Ctenopteryx Appellof Family 3. Enoploteuthidae Chun, 1910 Pyroteuthis Hoyle, Ptery gio; teuthis Fischer, Lycoteuthis Pfeff., Lampadioteuthis Berry, Nemato- lampas Berry, Enoploteuthis Fer. and Orb., Abralia Gray, Abraliopsis Joubin, Ancistrochirus Gray, Thelidioteuthis, Preff. Family 4. Onychoteuthidae Gray, 1849 Onychoteuthis Lichtenst., Ancistro- teuthis Gray, Deleoteuthis Verr., Chaunoteuthis Appellof, Tetronycho- teuthis Pfeff., Moroteuthis Verr. Family 5. Neoteuthidae nov. Neoteuthis nov.t Family 6. Octopodoteuthidae Berry, 1912 Octopodoteuthis Rupp. Cucioteuthis Strp. Kelaeno has no hooks. Hooks have not been found in any fossil teuthoid. My erroneous view was based on incorrect data of Munster, Meyer and Wagner, which appeared all the more credible as they were con- firmed by J.Walther in 1905. The error is based on a faulty determination: the corresponding plate of Daiting with the cone, proostracum, head impression, eyes and arms in the Munich study collection does not belong here (contrary to appearances) but to Acanthoteuthis (p.112). This error has led to my view (pp.127 —133) that the Teuthoidea and Belemnoidea evolved from a common ancestor with hooks. The hooks of Belemnoidea are not related to those of the recent Teuthoidea, although both have developed in a similar manner. This may be a special type of metateuthoid; in any case, it appears from the structure of the heart that it is not an oegopsid (cf. Chapter 4). Here also belong an undescribed species of oegopsid of which I have 3 young stages with the following des- cription: habitus Sepioteuthis-like; fins lateral, rounded; mantle sac slender; gladius with sharp posterior cone, otherwise Loligo-like; arms short, biserial; clubs quadriserial. 46 49 Family 7. Family 8. Family 9. Family 10. Family 11. Family 12. Family 13. Family 14. Suborder 3. Family 1. Family 2. Family 3. Histioteuthidae Verr., 1881 Histioteuthis Orb., Calliteuthis Verr., Meleagroteuthis Pfeff. Architeuthidae Pfeff., 1900 Architeuthis Strp. Brachioteuthidae Pfeff., 1900 Brachioteuthis Verr. Chiroteuthidae Gray, 1849 Chiroteuthis Orb., Mastigoteuthis Verr., Idioteuthis Sasaki, Joubinioteuthis Berry, 1920 Grimalditeuthidae Pfeff., 1900 Grimalditeuthis Joubin Cranchiidae Gray, 1849 Cranchia Leach, Pyrgopsis Rochebr., Liocranchia Pfeff., Liguriella Issel, Phasmatopsis Rochebr., Toxeuma Chun, Taonius Strp., Desmoteuthis Vier. Meersialioeiramyic hiita Pfefte, iavomudsinuimi Piet... Cie insitallwlioiteuitqhrs Chun, Phasmatoteuthis Pfeff.,Gali- teuthis Joubin, Corynomma Chun, Teuthowenia Chun, Bathothauma Chun, Verrillioteuthis Berry, 1916, Leucocranchia Joubin, 1912 Ommatostrephidae (Gill, 1871) Strp. MIMWVessy Stroy o.dia tops dish Garand Nie tion darus Pfeff.,. Ommatostrephis Orb., Dosidicus Strp., Hyaloteuthis Gray, stenoteuthis Verr.,Symplecto- teuthis Pfeff., Eucleoteuthis Berry Thysanoteuthidae Keferstein, 1866 Thysanoteuthis Troschel Sepioidea Naef, 1916* Belemnosidae Naef Belemnosis (differs from Spirulirostra) Belopteridae Naef Belopterina M.-Chalmas, Beloptera Blainv.. Belopteridium Naef (for Beloptera edwardsi Deshayes, 1825), Belopterella Naef (for Beloptera cylindrica von Koenen) Belospiellidae Naef Belosepiella Aless. * Protosepiodea and Metasepiodea have been omitted, since transitional forms have been found. 47 Family 4. Spirulirostridae Naef*t Spihrulvros tra Orb. Spinuli nos t- rella Naef (for Spirulirostra szainochae Wojcik, 1903) Family 5. Spirulirostrinidae Naef Spirulirostrina (transitional to Sepiidae) Family 6. Sepiidae Orb., 1845 Sepia L., Sepiella Strp., Hemisepius Strp., Metasepia Hoyle, Belosepiat Voltz Family 7. Spirulidae Owen, 1848 Spirula Lam. Family 8. Idiosepiidae Appellof, 1898 Idiosepius Strp. Family 9. Sepiolidae Tryon, 1879 Sepiadarium Strp., Sepioloidea Orb., Rossia Owen, Semirossia Strp., Heteroteuthis Gray, Nectoteuthis Verr., Iridoteuthis Naef, Stoloteu- this Verr., Sepiolina Naef, Euprym- na Strp., Sepiola Leach, Sepietta Naef, Rondeletia Naef** Order II. OCTOPODA Leach, 1818 Suborder 1. Palaeoctopoda nov. t Family 1. Palaeoctopodidae Dollo, 1912 Palaeoctopus Woodward, 1896 Suborder 2. Cirroteuthoidea Berry, 1920, nom. nov. (Pteroti Reinh. and Prosch 1846; Lioglossa Lititken, 1882; Cirrata Grimpe,1916) Family 1. Cirroteuthidae Keferstein, 1866 Cirroteuthis Eschricht, Stauroteu- this Verr., Frockenia Hoyle, Cirro- thauma Chun, Chunioteuthis Grimpe Family 2. Opisthoteuthidae Verrill, 1896 Opisthoteuthis Verr., 1883 Family 3. Vampyroteuthidae Thiele, 1915 Vampyroteuthis Chun, Melano- teuthis Joubin, Laetmoteuthis Berry, Hymenoteuthis Thiele * Belemnosis Edwards, originally placed in this family, is stated by F. A. Bather to be probably an eroded Spirulirostra. I accept this view in the sense that I consider the abberrant shell fragment to belong to a Spirulirostra-like form (Bullen- Harris, 1894). ** The name Rondeletia Naef,1916 is already occupied by a fish and should be changed to Ronde - letiola (Proc. U.S.Nat. Mus., Vol.17, p.454). 48 ol 52 Suborder 3. Polypodeidea nov. (Apteri Reinh. and Prosch, 1846; Trachy- glossa Lutken, 1882; Incirrata Grimpe, 1916) a. Ctenoglossa nov. Family 1. Amphitretidae Hoyle, 1886 Amphitretus Hoyle Family 2. Bolitaenidae Chun, 1911 Bolitaena Strp., Eledonella Verr., Vitreledonella (?) b. Heteroglossa nov. Family 1. Octopodidae Orb., 1845 Olctopus, Lam) Ss caeuricusmirosch:, Eledone Leach, Velodona Chun, Cistopus Gray, Pinnoctopus Orb., Graneledone (Joubin, 1918) Family 2. Argonautidae Cantraine, 1841 Argonauta L., Ocythoé Rafinesque, iremoctopus: Delle Chiaje; Alloposus Verr. d. SCHEME OF THE ORGANIZATION OF MOLLUSCA A detailed treatment of the original structure or typical organization of molluscs involves a discussion of a large number of data and is beyond the scope of this work. However,a general outline of the phylum Mollusca will be given. For didactical purposes, this will be based on a scheme which describes a number of morphological aspects without pretending to be a reconstruction of the ancestral form or of a methodically constructed type. This scheme represents the relationships in the subclass Eumalakia (Naef, 1911) (Cephalopoda, Gastropoda, Lamellibranchiata and Scaphopoda) more accurately than those in the Amphineura (Placophora, Solenogastres). It shows an undivided, strictly postvelar shell (21), of which the conical shell has to be considered as the primitive form. Sucha shell permits a gradual change from a flat or blunt form to various degrees of a pointed, long form with corresponding changes in the scheme. The apex of the shell (22) represents the primordial or embryonic shell, the origin of which will not be discussed here. The later form of the shell is the result of more or less uniform marginal growth by continued deposition of fresh shell substance on the free margin of the shell. This growth is evident in the growth lines of all typical shells of molluscs, in which the successive stages of develop- ment can be followed. The solidity of older shells, however, depends on growth in thickness, i. e., secondary deposition of new shell structure on the inside of the original plate. It produces a thick, lamellar strengthening layer (hypostracum or nacreous layer). The nonlamellar substance produced by 49 (51) FIGURE 5. Scheme of organization of molluscs: 1 —cerebral part; 2 — pleural part; 3 — pedal part of the esopha- geal ring; 4 —statocyst; 5 — foot; 6 — projections of the retractor of the cephalopodium or shell adductor; 7 — insertion of this muscle on the shell; 8 — anterior branchial ganglion; 9 — median point on the anterior transverse mantle groove; 10 — posterior part of mantle cavity; 11 — free margin of shell; 12 — free margin of mantle; 13 — mantle; 14 — posterior gill; 15 — anterior gill; 16 — position of anus in the middle of posterior transverse mantle groove; 17 — visceral ganglion; 18 — heart; 19 — hind intestine; 20 — epithelium of shell; 21 — shell; 22 — apex of shell (embry- onic shell); 23 —stomach; 24 — muscle of shell; 25 — esophagus; 26 — pleurovisceral cord; 27 — most anterior point in groove of mantle; 28 — anterior part of mantle cavity; 29 — eye; 30 — tongue; 31 — buccal cavity; 32 — snout; 33 — sac of radula; 34 — section through mantle margin with a weak fold of shell (36); 385 — section through mantle margin with a more strongly developed fold of shell. marginal growth is known as the ostracum or prismatic layer. Both layers consist of an organic substance (conchin) with deposits of calcium carbonate. The ostracum is naturally weakest at the apex (i.e., at the embryonic shell) and becomes gradually thicker towards the margin. The layers of the hypostracum (nacreous matter), on the other hand, grow in number and thickness from the margin to the apex. In many Eumollusca these two typical layers are covered by a third, the periostracum, which is formed by secondary deposits on the outer side of the ostracum (Figure 6, p. 53). 50 (53) FIGURE 6. Diagrammatic median section through the shell of a typical eumollusc: 1 — embryonic shell; 2 — periostracum; 3 — ostra- cum; 4 —hypostracum. The hypostracum becomes thicker toward the apex because the first formed nacreous layer does not reach the present margin of the shell. The ostracum, however, becomes thinner in this direction because the young animal built a weaker shell; a secondary thickening of the shell is not possible. The periostracum becomes thicker towards the free margin of the shell if the shell fold (which secretes it) overlaps only slightly (cf. Figure 5). The entire shell is formed by the epithelium of the shell (20), an epider- mal layer adjacent to the shell, but not grown together with it. The ostra- cum is produced by the outer layer of the epithelium, the hypostracum by its other part, by differentiated glands. The shell fold produces the perio- stracum. The marginal areas of the shell epithelium belong to a thin, annular skin fold which hangs free from the visceral mass and adheres to the shell. This is the primary or skin mantle (13). Between the skin mantle and the body lies the mantle cavity (10). The mantle cavity opens outward in the ring-shaped mantle slit. The mantle cavity is narrow anteriorly (28) and laterally and its deepest part is a simple ''mantle groove’ (27). It widens posteriorly so that the inner boundary of the mantle cavity forms a wide plane known as the roof of the mantle cavity (9-16). The anterior mantle groove (9) divides the roof of the mantle cavity from the body, the posterior groove (16) from the mantle; both grooves meet laterally in the simple mantle groove. The gills, or ctenidia (14, 15), are situated in the posterior part of the mantle cavity, attached to the roof of the mantle cavity. This "axis'' is an apically tapering plate (epithelial fold). The afferent vessel is situated at the median margin of the plate, the efferent vessel at the other margin. Between them there are secondary folds alternating at both sides of the 53 gill axis. These are the "lamellae of the gills.'' Each basis of the gill is accompanied by the opening of a coelomoduct (renal-genital opening). The above parts of the body, situated on one side of the line 3—27 in Figure 5, form a natural morphological entity which is connected with the shell and is named shell sac or mantle sac. It contains the main ''vegetative" organs. The other part of the body or cephalopodium contains the locomo- tory organs and the main neural and sensory centers (oesophageal ring (1,2, 3); statocysts (4) and eyes (29)). This is the ''animal"' part, which is functionally distinct from the ''vegetative'' part. The cephalopodium is often divided into head and foot (5), but we shall not attempt to determine whether 51 54 such a division is typical (cf. Naef, 1911, pp. 85—86). Head and foot are closely associated in typical Eumollusca and indistinguishable in many cases. The head bears a pair of invaginated (open) or vesicular eyes (29) and continues in a snout (32) in which the mouth opens. The typical structure of the foot part is very problematic. At any rate, 1 do not assume a simple creeping sole as the basic form for Eumollusca. A fuller development of locomotory-sensory appendages is more probable, but this does not appear in the scheme. The figure explains the inner structure and shows clearly the major as- pects of the typical organization. To be noted is the paired ''shell muscle" or cephalopodial retractor (24), its diffuse origin in the cephalopodium (6) and its insertion on the inner side of the shell(7). The insertion is formed by the shell epithelium, the characteristic transversely fibered structure of which at this point provides a firmer connection with the shell than could be achieved by mere adhesion. As regards the change or the ''metamorphosis" of the conditions shown in the scheme, we will mention only one point. The formation of the periostra- cum as defined above (p. 52) presupposes at least a partial folding of the mantle onto the outer side of the shell for some time. This is achieved by a special skin fold which originates at the mantle margin and is called the "shell fold.'' Such a fold may cover the shell to a varying extent and even enclose it completely, forming secondarily an inner shell. This condition, which is characteristic for all dibranchiate cephalopods, cannot be described simply as growth of the mantle around the shell. The mantle takes part in this process only in providing the shell fold. e. NORMAL ORIENTATION OF THE BODY The general orientation of the body as shown in the scheme permits recognition of anterior and posterior, dorsal and ventral parts. This condi- tion is characteristic for other molluscs, but in the normal position of the body of a cephalopod, the shell apex is directed more or less posteriorly and the cephalopodium anteriorly. The posterior side of the schemes thus becomes ventral, and the anterior side dorsal. The posterior gill then becomes the lower, the anterior, the upper gill. Similarly, the posterior mantle groove becomes the lower, etc. This natural (''physiological'') posi- tion of the body, usually assumed during swimming, is of basic importance for the distinction between upper, lower, anterior and posterior in the following text. Where another morphological orientation takes place, it will be stressed. This orientation should not be applied generally, as was done by Land and Hescheler (1900). This would mean ignoring the natural orientation of the body, and would cause confusion without morphological clarification. The topographic relationships in molluscs are much too complicated (e.g. torsion and detorsion, regulation of the shell position in gastropods; Naef, 1911) and variable for general orientation to be of any morphological help. Distinctions like anteriorly- posteriorly, ventral-dorsal or right left do not provide a significant comparison. A differentiated treatment of the problem of position is necessary, which is easy and natural in our case. 52 55 5. THE RECENT GENUS NAUTILUS AND THE FOSSIL TETRABRANCHIATA Contents: a. Nautilus. General form and shell. Suture lines. Form of soft parts. Mantle organs. — b. The fossil Tetrabranchiata. Conclusions. a. NAUTILUS Although the genus Nautilus is not represented in the Mediterranean fauna, its prominent position among the manifold forms of the class calls for a brief discussion of its characteristics and relationships to the other forms. The description is based almost entirely on my own data, some new and others confirming existing knowledge, especially the comprehensive description of Griffin (1903) of the soft parts and of Appelléf (1898) of the shell, with whom I generally agree, except in the morphological evaluation of some important points. \ st ncn ON ww % WAS ANNA, 1} eK RY Parvo nq hu y ONS FIGURE 7. Lateral view of Nautilus pompilius in swimming position (alf natural size). The shell is without umbilicus, and the free mantle margin is visible dorsally below the "black substance" above the hood. Attention is drawn to the enlargement of the dorsal arm sheaths, toward the hood, the arrangement of the remaining arms, of which 4 are smaller than the others, and the ophthalmic tentacles, the displacement of the eye by the arms and their position with respect to the funnel. 33 56 Material for the studies on Nautilus was obtained from Pacific islands, including a collection made by Professor Dahl in the 1890s from Ralum (Bismarck Archipelago). Other material, brought by a missionary to Europe and obtained by the station, at Naples, was rather abundant (about 20 speci- mens) and consisted of immature half-grown specimens which provided some data forembryonic development. The genital organs of the youngest forms, especially the oviduct and nidamental glands, were at a stage passed in Dibranchiata soon after hatching; the gland lamellae, for example, were quite rudimentary and consisted of flat, ridge- shaped elevations (see Chapter 3). The material had been preserved for about 20 years in alcohol, and was not in ideal condition. However, this did not prevent determination of the major relationships, which is the object of this discussion. Jatta had already dissected some of the specimens, all of whichwereN.pompilius L. Results of the examination of this material will be published in detail elsewhere; this presentation will be confined to morphological elements, providing data for the determination of the typical cephalopod structure and allowing a comparison with the morphology of the Dibranchiata. As in all shell-bearing molluscs, the general habitus of the animal is largely determined by the shell which is well known. In recent species of Nautilus the shell may have a wide or narrow umbilicus or be without an umbilicus, and the last whorl of the shell in the mature animal may com- pletely envelop the preceding whorl (N. pompilius), or leave the axis of the coil free on each side (N. macromphalus and N. umbilicatus). The latter condition is undoubtedly the more primitive or typical, since it is also observed in young forms of N. pompilius. In the earliest parts of the shell, the coiling is even looser and the innermost whorl is completely evolute (Figure 9). The youngest stages of Nautilus have a horn-shaped, curved shell, resembling the shell of fossil forms like Gyroceras, Cyrtoceras and Orthoceras (Figure 8). (57) FIGURE 8. Three young stages of Nautilus (hypothetical) (natural size). They are the nucleus of older shells of 3 stages of growth; the soft parts of the young animal or embryo are only indicated. A freshly-hatched Nauti- lus has 7 chambers, the 8th chamber being much smaller than the preceding ones, as in young Sepia (q.v.). The successive stages have one, two, and three chambers. At the latest after completion of the third chamber, the dorsal margin of the shell must reach the site of the assumed embryonic chamber (indicated by dotted line), which must be removed to permit the second coil to be attached to the first. The first coil is evolute, of Cyrto- ceras type, the second coil is attached tangentially to the first, and the next coils show an increasingly involute nature: 1 — hypothetical embryonic chamber; 2 — first air chamber; 3 — second air chamber; 4 — third air chamber; 5 — living chamber; 6 — central hole through Nautilus shell resulting from the initially evolute condition and occupying the axis of the coil. 54 (58) say Fee eeRa ered SIS sj) FIGURE 9. Diagrammatic median section through Nautilus (a) and Sepia (b) fora comparison of the organization of Tetrabranchiata and Dibranchiata; Lm — lip membranes (outer and inner lip); Ok — upper jaw; Zg — tongue; Kk — hood; Nk — collar; Ml — dorsal lobe of the mantle; Sch — shell; Si — siphuncle; Sw — septa; Gd — poison gland (opening); Gg- cerebral ganglion; Pg — pedal ganglion; Vg — pleuro- visceral ganglion; Vd — esophagus; Vc — vena cava; Ed — intestine; Mg — stomach; Bs — caecum; Ov — ovary; Lh — body cavity (coelom); Gs — genital septum; Ps — peri- cardial septum; Hz — heart; Ni — kidneys; Km — gills; Mt — mantle; Int — sheaths of the inner series of arms; Ext — sheaths of the outer series of arms; Ci — cirri; Ar — arms; Tk — tentacle club; Bt — buccal funnel; Tb — ink sac; Ts — testis; Rs — rostrum; Os — osphradium; Nd — nidamental gland; Org — subradular organ; Gl — gastrogenital ligament. The coiling of the shell leaves a small empty porus in the axis of the coil between the umbilicus of one side and the other. The central part of this porus persists even when the openings are later closed. The Nautilus shell does not contain a typical central embryonic chamber. Since sucha chamber has not been found in larger specimens of the straight- shelled fossil nautilids, its absence has long been regarded as a specific character Sp) DU 58 59 of the Nautiloidea in contrast to the Ammonoidea. A typical embryonic chamber was demonstrated in young Orthoceras by Branco (1879), Clarke (1893) and Pocta (1902), while it was assumed that in Nautilus this chamber is later destroyed or lost (Broili- Zittel, p. 495), and this appeared quite correct to me at first. The shell nucleus is less curved than the following chambers, in contrast to most ammonites. On completion of the first coil, the shell margin should touch upon the normally developed embryonic chamber or its hypothetical site. If the animal grows further, this chamber must be removed. The second coil would then press against the first septum, which forms the posterior wall of the shell at this stage. The siphuncle must be retracted from the embryonic chamber (as in Orthoceras), and close the canal secondarily, probably where it pierces the first septum. A conspicuous linear depression is in fact always evident on the outer side of the shell at this point, which according to Hyatt (1893) is a scar (cicatrix) indicating the position of the embryonic chamber (Broili- Zittel, Figure 1,075, p.495). A more recent investigation led me to a dif- ferent interpretation: the scar, which is also present in many fossil nautilids, merely indicates the growth center of the shell, i.e. the site of the primary "anlage'' which grows later concentrically; the first chamber in Nautilus is actually the slightly modified embryonic chamber which is flat and plate- shaped rather than vesicular in form. This makes an interpretation of its secondary removal unnecessary (Figure 8). Other fossil nautilids develop in a Similar manner.* The formation of the septa of the shell should be interpreted as a modi- fication of the growth in thickness of other mollusc shells. The septa develop from the hypostracum (p.52) and merge directly with the more distal thickening layers of the shell. The septal necks are derived from the septa themselves, and the first two septal necks, at least, form septal caeca which envelop the siphuncle completely. In Nautilus,as in many fossil forms, only the first septal neck is solid; the following necks are calcified only near the septa (‘calcareous’), while the next parts are well developed but chitinous and enveloped by a weakly calcified, loose sheath. ** The sac-shaped body or mantie sac of the animal is situated in the "living chamber,'' so occupying it that the ventral and lateral mantle margins adhere to the inner side of the free shell margin. Above the shell opening of the living extended animal, the dorsal mantle margin covers an area covered by the ''black substance'' which represents morphologically the dorsal wall of the shell that adheres to the preceding coil. Accordingly, the upper boundary of the black zone corresponds to the dorsal shell margin which is covered. (Compare similar relationships in snails, for example, Helix pomatia.) The animal is attached only to a part of the living chamber. The concrescence of the epidermis with the innermost layer of the shell is restricted to the mantle margin and the annulus, a girdle-shaped zone located behind the origin of the mantle fold which is narrow in its upper and lower * After having drawn the above figure, I obtained shells of young fossil nautilids (Cyrtoceras, Gyroceras, Trochoceras, Nautilus, Lituites) which closely resemble recent Nautilus in having no initial vesicular chamber, a siphuncle beginning close to the apex, and a relatively straight initial part of the shell. Other nautiloids differ considerably from this and I intend to use this syste- matically. * A detailed description of the structure and development of the Nautilus shell will be found in a syn- thetic work on fossil cephalopods, which I intend to publish in the near future. In the present book, further details on the subject are only accessory and are outside the scope of this work. 56 60 (59) parts and widens considerably anteriorly and posteriorly at the sides. On empty shells this zone can be recognized by the chitinous deposit which makes adhesion possible; if this deposit is removed its position is indicated by a lack of sheen, on contrast to the adjacent areas. In the lateral part of the annulus, especially at its anterior widening, lies the origin of the retractor on each side. Behind the annulus is situated the last septum to which the body is attached except when a new septum is formed. The dividing line cor- responds rather exactly to the inner attachment of the septa, called the suture line. It should be borne in mind, however, that there are no muscle attach- ments on this line, the only one being that of the cephalopodial retractor, which borders on the anterior margin of the annulus and only extends posteriorly to the middle of the zone. The suture lines of the ammonites, whose typical (primitive) forms are closest to Nautilus, are not associated with muscle attachments, although the complex lines would result in a more perfect muscle attachment. The function of the suture lines is different and easy to understand. FIGURE 10. Lateral view of Nautilus pompilius. The shell is cut to expose the soft parts, and the last chambers are removed down to the median plane. Fleshy and hard parts of the siphuncle are visible: y — lower side of penultimate septum; x — upper side of penultimate septum, The last septum is still thin and incomplete. 1 — free dorsal margin of mantle; 2 — posterior boundary of suture line (annulus) which accompanies the attachment of the last septum to the shell wall; 3 — lateral lobe of annulus; 4 — saddle; 5 — posterior wall of the abdominal sac adjacent to the last septum; gs — rudiment of the genital septum; 6 — anterior margin of suture line; 7 — anterior margin of the area of attachment of retractor (8); 9 — dorsomedian border of the mantle sac pressed in by preceding coils of the shell (10); 11 —narrow dorsal part of the annulus; 12 — free margin of shell; 13 — funnel pocket; 14 — oph- thalmic tentacles; 15 — hood; 16 — eye; 17 — part of hood produced by the second pair of arms; 18 — funnel; 19 — tactile arms; 20 — bor- der of the black substance (dorsal margin of shell); 21 — ventral margin of shell; 22 — grasping arms; 23 — dorsal margin of shell; pd, — upper pericardial gland; pd, — lower pericardial gland; pce — anterior boun- dary of the pericardium and origin of the mantle; kb — branchial at- tachment; kv — branchial vein; km, — upper gill; km, — lower gill. The figure shows the relationships between animal and shell. Dd 61 The strength of the shell depends on the septa, and consequently the safeguarding of the animal against mechanical injury, including high water pressure at greater depths*; the septa also divide the shell in an appro- priate manner so that when a chamber is opened by piercing or a break, the animal is not exposed to harmful factors (small organisms and bacteria) nor is the static condition of the shell impaired. The simple septa strengthen the shell wall effectively but not uniformly, because the areas between the suture lines are left unprotected. A com- plicated winding, branched suture line would distribute the strengthening effect of the septum over a greater area and uniformly, and also increase the elasticity of the septum. Such a structure improves the resistance of the entire septum to impact and achieves the best possible results with minimal material. Another possible way of strengthening the chamber structure is by increasing the number of septa or thicker walls, which evidently would involve an increase of material and an impaired static condition which would have to be corrected (see the parts on Sepia shells in Chapter 27). The mantle fold is largely free and can be contracted or expanded while gliding over the shell surface. However, under natural conditions and in healthy animals the mantle fold generally adheres to the shell surface, so that no space is left for small enemies to settle and the outer mantle surface or shell epithelium can function for the thickening of the shell (p. 52). Let us now consider the freely projecting part of the soft body — the cephalopodium. Lateral examination shows a large eye the lateral surface of which represents an oval disk, slightly indented on the lower side and with projecting margin. The contractile ''pupil’’ in the middle is an opening which connects the eye chamber with the water. A groove (ciliated groove) extending from the pupil to the marginal indentation probably serves to remove Slime and foreign bodies from the eye. This disk lies on the eyeball which is situated on a thin stalk (Chapter 2) and is inserted in a depression formed by parts of the arms; in this manner, the arm bases and the eye form a more or less uniform surface with the shell margin. Posteriorly and below, the eye depression borders on the funnel pocket, which is almost completely covered by the shell margin in Figures 7 and 10. Below the eye lies the olfactory organ, the rhinophore, projecting externally in a short, tentaclelike process with the openings of two deep pits. The smaller, simpler pit opens dorsally at the base of the process, the other pit is deeper and opens on the process itself, piercing it to the apex and penetrating deep into the head, almost to the statocyst; this pit is lined with sensory epi- thelium. Only embryological studies can show whether this complex struc- ture is entirely or partly homologous to its simple counterpart in Dibran- chiata. There is possibly also a genetic relationship with the statocyst. The arm crown is situated in front of, below and above the eye. All the arms are uniform and consist of a proximal sheath and a distal cirrus (Figure 11). The cirrus is a finely annulated, mobile process, resembling an earthworm with one end inserted deep into the sheath and the other more or less protruding. Morphologically, the sheath is a differentiated arm base which envelops and protects the delicate distal part of the arm. * All Nautilus shells show greater or lesser regeneration from injuries of the living chamber. This is a characteristic of a littoral mode of life. 58 (63) The cirrus can be greatly extended and shortened; it can be retracted almost completely into the sheath or project in its greater part. This specialization of the arm base into a sheath is specific for Nau- tilus. The Dibranchiata show no trace of it, which suggests that sheaths were absent in the common ancestors of these groups (Chapter 1). By growing together into complexes of various size, the sheaths make mor- phological orientation within the arm crown extremely difficult. In these complexes only the distal end with the cirrus opening remains free; the true arm base, including the base of the cirrus, remains concealed. Never- theless, the following may be recognized (Figure 12, p.65). FIGURE 11. Cirri of Nautilus pompilius: a — cirrus of the outer series of arms (grasping arms) (*4 natural size); a; — same, highly mag- nified in profile; a)—transition between the inner and outer surface; a, — profile of the proximal part; a4 — cross section through the distal part; b — sensory cirrus (CH natural size), inner and lateral aspect,comblike; b, — cross section; c — ophthalmic cirrus (apex) (Cie natural size); cy, — cross section. The cross sections show adhesive pads on the arms within which the oval cross section of the nerve is visible, and the artery and the vein (black) below. Note the variable form of the adhesive pads. 59 62 Two very short arms, called ophthalmic tentacles, with short sheaths are situated above each eye, one before it, the other behind it. In front of and below each eye, markedly behind the other arms, are another four arms. These 6 arm pairs differ structurally from the other arms, especially with respect to the cirrus. They are incapable of grasping, and are mainly sensory organs; I have called them ''tactile arms.'' An outer series of strong arms, 15 on each side, surrounds the mouth. The bases of these arms are not arranged in a simple circle, but form a zigzag pattern, so that the outer series can be divided into an inner and an outer row. (In the schematic drawing of Figure 12, this arrangement has been simplified and the number of areas arbitrarily reduced to 6.) There is an inner series of arms which surrounds the mouth. The inner arms, or oral arms, are smaller and more numerous. There are up to 32 on one side in the female and their sheaths are fused into three muscular lobes, two lateral and one ventral. In the male, the ventral lobe is rudimentary or converted into the organ of Van der Hoeven (Figure 12b), and the ventral parts of the lateral lobes (each with 4 cirri) are separated and strongly modified into copulatory organs (‘'spadix'' and ''anti-spadix''). The two mediodorsal arm sheaths of the outer series are markedly enlarged, with a tough, warty skin; they extend posteriorly to the dorsal mantle fold to which they adhere over a wide area. Anteriorly they are joined by the second pair of sheaths with a similar surface, con- tributing to the formation of the hood which can close the entrance of the shell when the animal has retracted inside, like the operculum in snails. (In ammonites the "aptychus'' apparently serves to strengthen the hood.) In the quietly swimming animal (Figure 7), the hood protects the animal at least from above. Contraction of the dorsal mantle lobe brings the hood (i.e. the dorsal arms) in direct contact with the shell, particularly with the FIGURE 12. Diagram of oral area and arm apparatus of Nautilus pompilius: a — female; b — male. 1 — upper jaw; 2 — lower jaw; 3 —imner lip; 4 — outer lip; 5 — lateral lobe; 6 — medio-ventral lobe of the inner arm series; 7 — organ of Valenciennes; 8 — receptor for the spermato- phore; 9 — cephalic shield passing laterally into the rest of the cephalic sheath, i.e. the product of fusion of the sheaths of all 30 arms; 10 — ophthalmic arms; 11 — tactile arms; 12 — funnel; 13 — organ of Van der Hoeven; 14 — anti-spadix; 15 — spadix. 60 63 64 black substance (see above, p.59). However, this does not prove that this shell layer is produced by the arms* (cf. Figure 9). The shell epithelium, i.e. the outer lining of the mantle sac, produces the entire shell. As in all typical molluscs, shell growth is marginal, made possible by the free mantle margin, i.e. the peripheral part of the shell epithelium. The black substance, representing the covered dorsal part of the shell margin, is secreted by the dorsal lobe of the mantle. All the sheaths of the outer series of arms, including the hood, are fused basally into a head sheath (Figure 12) inside which they show a biserial, zigzag arrangement. The arms of the outer series and sheaths of the adjacent tactile arms (Figure 7), separated by grooves, project above the surface and are curved elbowlike posteriorly, providing protection for the eye. In the female, the inner medioventral part of the head sheath is differentiated to accept the spermatophore during copulation. It has a plicate, glandular surface (Figure 12). Thus, the arm series has important genital functions in both sexes, especially during copulation which proceeds with intertwined arms, mouth against mouth, as in Sepia and many other decapods. The cirri of the outer series of arms have a typical structure. The base, always contained in the sheath, is almost cylindrical and passes gradually into the extensible part which has a more or less flat, unpigmented inner surface, facing the mouth and a rounded, brown outer side. The cirrus has fine rings, except at the base where the sculpture is indistinct. Proximally the rings are very dense and rather uniform; distally they become in- creasingly wider and on the inner side assume the specific structure necessary for grasping. The incisions become much deeper here and the dense, raised edges form a flattened outer surface. The interior of the prominent parts has a specific musculature which extends from the flat adhesive surface to the rigid, muscular axis of the cirrus, converting each ring into an incomplete sucker or adhesive pad. In view of the large number of rings — about 30 on the projecting part of the cirrus — the combined effect of 30 such cirri in the outer series is considerable. Furthermore, in the largest and best preserved specimens there is always a median groove along the distal part of the prehensile surface of the cirrus. Since the separate pads are closely contiguous and are connected by mucus, the cirrus as a whole can function as a sucker and increase adhesion. The cirrus musculature is similar to that of the arms of the Dibranchiata, especially in the well defined axial part. The same is true of the nervous and vascular supply of the cirri, as well as their orientation around the head. There is no doubt that each cirrus corresponds to a dibranchiate arm. The adhesive rings are homologous to suckers. This is evident from the embryological development of the suckers (cf. next chapters and Volume II). The main function of the outer arms, i.e. of the cirri of the outer series, is to capture the prey, which is then transferred to the oralarms. Nau- tilus feeds mainly on fish; the crops of the specimens I dissected were packed with remains of a sardinelike fish, partly in large pieces. Ido not * G.Steinmann (1890) maintains this without proof. He also states that the shell of Argonauta is produced first by the mantle, and only later by the arms. By such irresponsible statements formulated as actual observations, Steinmann attempts to prove his abstruse view on the origin of Argonauta from the Tetrabranchiata (ammonites) (cf. Chapter 52). 61 65 66 know whether the animal is agile enough to capture such fish or only slow moving species or dead fish. Nautilus was caught with dead fish as bait in weir baskets, so that all specimens had a last meal of this kind, and the stomach contents reveal little about their normal diet. In my view, Nau- tilus uses its cirri to capture small, benthic animals in their hiding places; such an activity is compatible with the limited grasping power of the arms. However, these arms also serve for creeping. Willey observed that the animal can adhere to smooth surfaces by means of the cirri(Lang-Hescheler, p. 282): The oral arms are similar in structure to the outer arms, provided they are not secondarily modified into genital or other organs (organ of Van der Hoeven, Figure 12), and they are therefore also adhesive organs. The tactile arms, especially their cirri, are markedly different in structure. The ophthalmic cirri are ciliated in contrast to the other arms (?). Their adhesive pads, however, are highly modified; they surround the cirrus in the form of sickle-shaped lamellae, leaving only the outer side free (Figure llc, ci)and their widest part projects markedly from the cirrus, cre- ating a comblike profile (c). The ophthalmic cirri are very short and taper apically more sharply than the other cirri. In addition to their function as sensory organs, the ophthalmic cirri clean the eye socket. The four tactile arms (p.62) are also unsuitable for grasping. Their adhesive pads do not occupy the entire width of the cirrus, although they project enough to give a comb effect (Figure 11b). The lamellae of the organ of Valenciennes (Figure 12a, N:7) represent modified cirri and often show transitional forms to the adjacent cirri of the inner series. The oral arms surround the mouth, which is surrounded by a wide circular fold or outer lip with rather thin edges, like in the Dibranchiata. There usually projects from the outer lip also a second circular fold, the ''inner lip" with a thick margin covered with papillae. The outer lip corresponds to the boundary of the primary, embryonic mouth,* whereas the inner lip originates at the interior of the stomodeum and belongs to the anterior part of the buccal cavity. The inner lip usually covers only partly the biting edges of the jaws, and the tongue with the radula, is usually situated on the base of the buccal cavity and can project through the open mouth, so that these parts can be considered as external, although they are described here for practical systematic reasons. All these relationships are summarized in Figure 12. The jaws (Figure 13)are large cuticular structures developing from two muscular processes at the entrance of the oral cavity behind the inner lip. The lower jaw projects anteriorly and laterally beyond the upper jaw, so that they act like scissors in biting. Each jaw consists of an outer plate (A) and an inner plate (J), which meet in the angular, protruding biting edge (B). When partly open, the whole structure resembles an inverted parrot beak. Both jaws of Nautilus are covered apically and at the biting edge witha calcareous white substance which is denticulate on the lower jaw. The rest of the jaws consist of chitinous material. The inner plate of the upper jaw is much stronger developed than the outer plate and projects far beyond this; in the lower jaw the proportions are reversed, particularly in Nau- tilus where the inner plate is completely covered in side view. * Easily observed in the development of Dibranchiata. 62 67 63 J As in other molluscs, the radula consists of a bandlike, curved ''main plate'’ covered with dense transverse rows of chitinous denticles. Its posterior end is situated in the radular sac (Figure 9, Zg) where denticles are continually developed as the radula moves forward. The = anterior end extends above the tongue where the —¥ retrorse teeth widen into a compact brush. This L part is strengthened by the "lateral plates’ which are connected with the main plate and cover the tongue on both sides. As the animal grows, the radular elements increase in size successively and the radula extends backward at a rate propor- tional to general growth and inversely proportional to the wear at the anterior part. The arrangement of the dental rows closely resembles that of typical Dibranchiata. On each side there is a marginal row of flat, transverse, a — upper jaw with inner marginal platelets without distinct cusps. Next plate J, outer plate A, biting is a row of long, subulate, curved, outer brush edge,B, palatine maygin L, cal- teeth which cover an intercalated row of flat, careous coat K; b — lower : 5 neat transverse teeth (intercalary platelets) serving jaw (same). Note the denti- FIGURE 13. Jaws ofNautilus pompilius (natural size): tion (Z) along the biting as support, but otherwise without function (this edge (B). Main figures,lateral; row is invariably absent in the Dibranchiata). small figures, inner view. Toward the center of the radula there is a row of "inner brush teeth" and further rows of small, short rasping teeth, one paramedian and one sub- median. Finally, there is a median row of rasping teeth which project slightly above the preceding rows. (The paramedian row of rasping teeth is absent in Dibranchiata.) Thus Nautilus has 13 rows of teeth compared with 9 rows in typical Dibranchiata. Except for the flat teeth, a radular consists of a basal plate inserted in the main plate of the radula and of a dental process with cusps which extend into the biting edges (cf. Tables 14—16). The funnel apparatus is strongly developed in Nautilus and some pelagic nektonic dibranchiates in connection with its function. The funnel consists of a muscular lobe which originates behind the bases of the ventral arm sheaths and forms a large tapering sac with the apex pointing outward. When water is expelled from the mantle cavity, the resulting pressure presses the ventral posterior margin of the funnel to the mantle and shell margin (Figure 9, p. 58), closing it completely. A valve on each side of the mantle slit prevents water escaping into the funnel, closing the opening. This is the funnel pocket. The mantle slit (Figure 7) is closed by the collar (Nk) — a membranous lobe which is directed forward and is situated between the cephalic shield and the dorsal lobe of the mantle, adhering to both by its smooth surfaces. (This is homologous to the neck cartilage of the Decapoda. ) 63 “syeays wae — 1y ‘eka — ny ‘leyood youuny _—1p {jouunj — 1p {eavo euaA — A ‘JOIORNeI youUN} — TY OWEN yetpodoreydao — oy ‘squawyoenie [11s Suveq SuITTOAS — M ‘dTOISoA wuojt1d yo Sutuedo yeIuas (je) eTewW — q ‘Sutuado yeiTues (AYysts) gyeuiay — AQ ‘snue — e ‘Te Aaupry ou ysnoity o[qista sesepuad -de urea — eA ‘Aaupry seddn — 4n ‘Aoupry saMoy — 'N ‘aiod [rudd yoddn — %u ‘o10d yeuar JaMoy — Tu ‘uTea TeryouRriq — AY ‘(ain3ty ut poyiew ou) wintperydso yaddn — %@Q ‘umtpeasydso ramoy — 'O ‘purys [equowieptu — pn ‘sqqz3 eddn pue 1amoy — @y pue ty ‘apueU — WW 1(@ZISs [PINIeU Ford pepploy xoTduoa([euTMOpqe) [Rue ot SUIpN[oUT ‘Trays ou mnowIM snipidurod snptineN joApog ros “gt FYNOls (67) -eleIYOURIGIG JO soTstiaoeIeYyo yeotdéy aul [Te sMoys OBIT OT ‘astMIeyIO ‘OB1[ 0] UI 1OUTISTP ssoT ST Yaoi Burdses pue ysniq UaeMieq ise -uo0d oui pur A[MoO]S SUaPIM OST[O7T Jo 1eU) ‘(SoIeTYyoURIG -1p Auewl ut se) Appayieu suapIm snplineN jo e[npel SUL “JOPIUTST SN[IINEN Ul MOI asIoAsueN IaMOT PUOdeS am ATU ‘pasnun st OST{O7T JO 1eYI S[TYM ‘UIOM ST SNT -11N BN jo uMoys wed ay, “RIeTyoURIqIq pue vielyoUurIq eile J, ueemiag uostieduioo [eioues & Iuled ssuIMeIP oUL -aiejd peioiey — 6 {yjooi Burdses uelpaut — g ‘Yiee1 sutdsei ueipawigns — 1, :yj001 Surdses uelpaureied — g ‘yie01 YstTuq youul— @ ‘slopeietd Aiepeoisiul — p ‘yjoo1 Ysniq Jeno — ¢ ‘syaqaieyd yeulsiew — z ‘oietd leur oui jo uTsieW — T :poudiey oie seynpel eUL “(x6 Pry -lU3eW) (q) 1S9q10} OBT{OT puke (xg'9 polluseul) (e) snijidwod sn{1ineN jo e[npel jo sued “pT AYNold 64 69 70 In the umbilical region of the shell, the collar passes into the funnel sac where the mantle cavity is very narrow and shallow. The funnel pocket and funnel form together a large, saclike organ. We assume that this organ can actively expand and receive a large volume of water through the aper- ture between the mantle margin and its own free edges. As the whole sac contracts from the front, the free edges are pressed against the mantle and the water is ejected out of the funnel. The resultant repulsion provides energy for movement in a direction opposite to the opening (see Chapter 1). Another factor comes into play for more powerful propulsion. By the contraction of the cephalopodial retractors the whole anterior part of the body is retracted to some extent into the living chamber and this reduces the volume of the mantle cavity behind the inner opening of the funnel (and that of the funnel itself). A third possibility of obtaining locomotory energy appears first in the Dibranchiata and consists of contraction of the mantle, which in Nautilus is thin and membranous and adheres to the inner side of the shell (cf. Chapter 2). The intake of water into the funnel and mantle cavity activates another valve situated in the distal part of the funnel. This funnel valve (Figure 9, Tk) closes the funnel opening during inhalation and regulates the intake of water through the mantle slit. The base of the mantle cavity is of particular morphological interest. While it is very narrow dorsally and closed by the collar and still quite shallow laterally in the umbilical area, the cavity becomes much wider and deeper ventrally (Figures 9 and 10). A posterior lower plane bordering on the mantle cavity corresponds to the roof of the mantle cavity in the molluscan scheme (p.52). Attached to this is typically a complex of organs folding upon the body just like the mantle, so that its parts become situated ventrally and behind the deepest part of the mantle cavity (Naef, 1913, p. 387, etc.). I called this area at first the anal complex, but later changed this to abdominal complex, which is more suitable and can be better applied to homologous parts of the Dibranchiata. The abdominal complex can be folded back together with the mantle (Figure 15) or separated from the deepest mantle slit as shown in Figure 16. Naturally, what appears here to be transversely extended is, in fact, curved together into a horseshoe. The abdominal complex is curved back like the mantle, but this condition which I have named "retroflexion'' (Naef, 1913, p.86) should not mislead us into thinking (as Griffin did) that the abdominal complex belongs directly to the mantle and that its organs lie within the mantle. The mantle is actually a clearly defined morphological entity, while the abdominal complex contains numerous organs and its connection with the mantle is only topo- graphical. The same is true of the position of the kidneys, gill bases and other organs in snails, mussels and Chiton with respect to the mantle. The upper mantle groove (Figures 15 and 16) is not straight but curved in a wide arc on each side before the large points of attachment of the cephalo- podial retractors. The part of the groove situated between these points is almost straight; anteriorly and upward it delimits the overhanging complex. Interrupted by the gill bases, the lower mantle groove runs downward and separates the abdominal complex from the mantle. The gills originate in this groove, each receiving its afferent branchial vessel from the abdominal complex (Ka) and returning an efferent branchial vessel to it (Kv). There is a close relationship between the gills and the abdominal complex. 65 A fleshy fold passes laterally and upward along the mantle groove, from the base of the outer side of each gill (Figure 16). On each side, both folds are fused into a flat ridge on the mantle (Figures 15 and 16W). attachments displace the gill bases mechanically to the side below the attachment of the cephalopodial retractors. Embryological studies of Dibranchiata and comparison with other molluscs show that the branchial attachments are not primary parts of the gills. In some young stages of Dibranchiata the branchial attachments resemble those of Nautilus and only attach the gill base to the mantle; this leaves the whole distal part free, although later the attachments grow along almost the whole organ. In still earlier embryonic stages, the branchial attachments are absent and the gill is directly based in the embryonic part corresponding to the abdominal complex without connection with the mantle (cf. Plate XVII and Vol. II). These branchial SAA ) : MI OD) FIGURE 16. Separated abdominal complex of a young female of Nau- tilus pompilius, flattened Gh natural size). The upper margin (cut) co- incides exactly with the oral mantle groove x, x,. The lower mantle groove follows the designations kb,, Ny, ptg, kby, No. rt — attachment of the cephalopodial rectractors; g — genital opening; g, — opening of pyriform vesicle; n — branchial nerve; a — anus; N,,N,— upper and lower renal sacs; pt, — area with opening of gonoduct; pt, — area with opening of pericardial funnel; kb, — root of the upper branchial vein, near the upper renal pore; kb, — root of the lower branchial vein, near the lower renal pore; acc — opening of the accessory glands; nid — nidamental gland area; os — lower osphradium; M — mantle; K, — upper gill; K, — lower gill; x —x, — upper mantle groove. The anus, surrounded by characteristic wrinkles and papillae, is situated in the middle of the abdominal complex. I assume that this is not its pri- mary position which it occupies in many other molluscs (C hiton, Gastro- poda, Lamellibranchiata), at least at first. in these forms (16 in Figure 5, p. 51), the anus is situated close to the origin of the mantle or the intestine passes secondarily for some distance along it. In predecessors of cephalo- pods and embryos of Nautilus, the anus probably lies in the middle between N2— Ne in Figure 16 and was then displaced toward the upper mantle groove. Beginning in Nautilus, this topographical displacement is further continued This may be assumed to be an important step toward {ial in the Dibranchiata. 66 72 an orientation in the complex of organs since it distinguishes morphologically between anterior and posterior parts in comparison with the typical Bila- teria. The intestine passes close to the surface on a stripe (Figure 16a) extending from the anus to the upper mantle groove. This stripe should actually extend to the lower mantle groove, and would mark the morpho- logical longitudinal axis of the complex, as it is only the retroflexed posterior end of the animal (cf. Naef, 1911, p.85, Figure 3; 1913, p.282, Figure 6). For the following discussion we shall orientate the complex morphologically, to facilitate comparison with other molluscs (Figure 5). ‘Anterior’ is then what was usually ''upper,'' while ''posterior" is 'lower'' in the sense of physiological orientation (p. 54). The surface of the abdominal complex or roof of the mantle cavity (p. 52), visible in Figures 15 and 16, is of particular morphological interest not only in Cephalopoda, but also for comparison with other molluscs (this will be discussed later). This surface is divided into a number of areas determined by the position of the different organs situated below the surface. On each side there is an area marked N2 corresponding to the larger, posterior renal sac, and an area Ni occupied by the smaller anterior renal sac; kbz cor- responds to the basal swelling of the posterior branchial vein, kb; to that of the anterior branchial vein. Both areas are joined by a third (marked pt2 and pti) at an angle and bearing an opening, the actual or morphological value* of which is that of a coelomic opening. The posterior pair of these openings is situated close to the renal openings and even united with them on common papillae in certain conditions of contraction. This pair of openings connects the pericardium directly with the mantle cavity, the pericardial funnel (Naef, 1913). The anterior pair differs on each side and is distant from the corresponding renal opening on the left and right. On the left side (right in the figures), a small opening situated in the mantle groove, slightly inward and above the renal porus, leads to the piriform vesicle; it is considered generally as a rudimentary gonoduct. Correspondingly, on the right, lies the female genital opening which in the adult animal forms a wide slit covered with characteristic concentric wrinkles on a large papilla, and continues deep into the opening to form the oviduct gland which produces the egg shell. The wrinkles probably form the inner layers of the oviduct gland, while outer layers develop from the nidamental gland, as in the Dibranchiata. I showed earlier (Naef, 1913) that both the functional and rudimentary genital openings represent true coelomic ducts and are homonomous with the openings of the pericardial funnel. These openings are not only situated close to the kidneys, ctenidia, branchial vessels and a number of visceral organs, but correspond metamerically (i.e. typically similar) to a number of internal organs; the regularity of this position is impaired only by the longer distance from the renal pores. It was assumed (Lang-Hescheler, pp. 359,341) and proven (Naef, 1913) that the pericardial funnels became independent by separation from the renal pores and that their primary opening led to the renal sacs near the pores, as in other molluscs. The gonoducts presumably moved some distance further in the same direction, and their openings were displaced away from the renal pores. An apparent * Morphological value: a structure is homologous to another showing the typical configuration of structures of this category; the second structure then indicates the morphological value of the first. 67 73 74 objection to this view (cf. Naef, 1913, Figure 12) is the presence of the mantle vein near the surface and between the anterior renal pore and genital openings. Embryological studies show that this vein moved here secondarily, so that there is no valid objection to the assumed displacement of the genital organs as indicated. (It will be shown elsewhere that there is also more significant evidence from other molluscs in favor of the supposi- tion of a primary metamerism or dimerism in the Mollusca.) In the male, there are ducts leading from the points corresponding to the genital openings of the female to a median papilla (position marked with X in Figure 15 on the vena cava v), which was erroneously considered as the penis. The ducts unite here into a single opening. On the right, the duct is directly connected with the gonoduct, forming a complete genital duct. On the left the duct is closed, without connection with the rudimentary genital opening. These distal parts are not comparable with the distal parts of the male genital ducts of Dibranchiata, and are special characters of Nautilus (cf. Naef, 1913, Figure 14, p. 406). The osphradia of Nautilus are characteristic papillae the function of which is perhaps (?) to check the quality of the inhaled water. There are two pairs of osphradia. One osphradium is situated on each side between the renal pores of the same side, forming a rounded lobule; the others form a group of 4 median elevations on the lower margin of the abdominal complex (Figure 16,0s). The lateral elevations of this lobe are probably metameric homonyms of the anterior osphradia. However, the morphological and physiological value of these structures is doubtful. The gills (Figure 17) resemble in principle those of Dibranchiata, and are symmetrical, at least in the ''anlage.'' They differ from the general form by extending transversely, in lacking a separate branchial gland in the fleshy axis, and in their strictly basal branchial attachments which are fleshy and not membranous as in the Dibranchiata. The afferent vessel (2) is situated very deep, while the efferent vessel (1) crosses the inner side of the gill; basally the vessels cross each other, creating the impression of a torsion (such a relation could be proven by comparison with Chiton and Gastro- poda! ). Between the afferent and efferent vessels lies the branchial axis, which is a thin plate at which the respiratory part of the branchial lamellae is attached, each lamella being situated between two lateral branches of the main vessels. The branchial lamellae alternate regularly, as in most molluscs, and are connected by skin folds with the fleshy axis. Each primary branchial lamella bears secondary lamellae in the form of ridgelike eleva- tions situated close together and arranged as alternating folds on each side of the lamella. Each of these ridges receives a deep afferent vessel and returns a superficial efferent vein. Alternating incisions divide the secondary lamella into tertiary parts with a similar vascular supply. A characteristic condition of Nautilus are the nidamental glands of the female, which appear late in life on the mantle area bordering on the abdo- minal complex. At first (Figure 16, nid) they represent a continuous system of low ridges extending from one side to the other. Later (as I observed in immature specimen), the marginal ridges form a fold which envelops the whole complex, forming an open pocket (Figure 15, Nd) which contains the glandular lamellar system of the original ridges. This large gland is un- paired, but the lateral parts of the pocket are much deeper than the median; if the median part becomes rudimentary, 2 sacs with glandular lamellae are formed and these are comparable to the nidamental glands of Dibranchiata. However, there is a marked contrast in position. 68 (73) FIGURE 17. Diagrammatic cross section through gills of Cephalopoda: a—Nautilus; b—Sepia; ec —Loligo. 1 — branchial vein; 2 — branchial artery; 3—ligament (of a single branchial lamella); 4 —merve; 5 and 6 — artery and vein of a branchial lamella. The dotted area in the lower part of b and c denotes the branchial gland situated in the solid "branchial axis" which passes into the branchial attachment. Characteristic of Nauti - lus (a) is the absence of a branchial gland in the fleshy axis, in which longitudinal muscles are clearly visible; in c, the perfora- tion of the branchial axis between 1 and 2. b. THE FOSSIL TETRABRANCHIATA The organization of the fossil Tetrabranchiata can only be discussed in close connection with Nautilus, but differently for the various groups. The primary organization of the Ammonoidea closely agrees with the above description, at least as regards the shell structure. The earliest ammonites are so Similar to Nautilus that they must be considered its closest 69 75 relatives, if not the ancestors, of the true Nautilidae. The ammonites with modified shells probably had a different soft body, although no specific conclusions can be drawn. At any rate, neither Nautilus nor the related Ammonoidea are directly connected with the origin (type) of the Dibranchiata. However, Nautilus belongs to the central group of Nautiloidea, although its position there is much less certain than its relationship to the typical ammonites. Genera like Orthoceras, Cyrtoceras andGyroceras certainly show at least the way in which the coiled Nautilus shell has evolved from the straight Orthoceras-like shell. Unfortunately, there is no direct information about the soft parts of these forms, and only indirectly, by a critical study of the correlation between shell form and soft body in Nautilus can some characters be interpreted as necessary adaptations or mechanical consequences of the shell form, while other characters have to be considered as an ancient inheritance or its metamor- phosis. The latter is especially true for characters shared by the Dibran- chiata which, in the last analysis, must have developed from the same root. These include peculiarities of organization not present in the Dibranchiata, but which must be assumed for their ancestors on morphological grounds (i.e. typical correlations and embryological data). Some secondary characters of the Nautilus form can be excluded inthis manner, first of all, for the straight Orthoceras-like ancestral forms of the Nautiloidea and also for the common ancestors of Tetrabranchiata and Di- branchiata (which may be identical or even more ancient than the former). This argumentation also eliminates as possible ancestors a number of fossil forms grouped around the genera Piloceras and Endoceras, since the structure of the siphuncle and other shell characters in these forms differ markedly from those of all other cephalopods. These two genera possibly (or probably) represent a more ancient type than the ancestral form of the other groups. FIGURE 18. Transition of straight Nautiloidea into the typical Nautilus form and reconstruction of the former from relationships in the latter. Four diagrammatic figures of young forms: a — Orthoceras-stage; b—Cyrtoceras stage; c —Gyroceras stage; d —Nautilus stage. The figure illustrates the modification of the soft body as a result of the coiling; these changes disappear gradually in the early stages. Instead of vesicular initial chambers, those in stages b and c must have been flat and bowl-shaped (pp. 57, 58). 70 The following characters are apparently secondary consequences of the specific shell structure in Nautilus: 1. the characteristic,dorsally impressed form of the posterior body inside the living chamber (Figure 10); 2. the formation of a special dorsal mantle lobe adjacent to the black substance (Figures 7 and 9); 3. the characteristic form of the collar (Figure 9); 4. the structure of the hood, at least in the final stage. (A certain protective function may be assumed for the upper arms already ina straight shell.) (Figure 7). These characters are the result of a secondary involution of the early coils into the later coil, the cavity and mouth of which encroach upon each other in the process. 76 On the basis of comparative studies of Dibranchiata, the following characters of Nautilus can be excluded as possible characteristics of the ancestral forms of the group: 1) the differentiation of the basal parts of the arms into tubular sheaths, which are completely absent in the Dibranchiata; 2) the specific complication of the olfactory organ, which in the Di- branchiata consists of a simple pad of sensory epithelium in the young stages or throughout life; 3) the characteristic division and differentiation of the inner series of arms in both sexes and, particularly, the formation of a ''spadix'’ and an "anti-spadix'' in the male; it is possible that some arms below the mouth in the female served as recipients for the spermatophores and that they were always transferred by the arms of the male; 4) the separation of the pericardial funnels from the posterior renal pores and the displacement of the genital openings from the anterior renal pores as well as the asymmetrical structure of the gonoducts (p. 71); 5) the formation of an unpaired secondary duct system connected with the primary genital openings and leading to the so-called penis (p. 72). If these specialized characters of recent Nautilus are ascribed toa common cephalopod type, an unnecessarily complicated path for the evolution of the Dibranchiata would result(see p.10). The phylogenetic approach would lead to a number of unnecessary hypotheses.* Such differentiations would then have to be cancelled and be replaced by others. These difficulties are avoided if the simplest, undifferentiated state is taken as starting point. By such considerations and assumptions, I obtained an idea of the type of relationships in the Cephalopoda as described in Chapter 1. In the follow- ing chapters these relationships will be examined in detail against the available evidence. * Influenced by the paleontological nature of the data concerned, abandoned the terminology of idealistic morphology without applying other methodical principles. Insofar as the systematic connection is con- cerned, I might as well have stated (p.74): “Neither Nautilus nor the Ammonoidea are direct ancestral stages (p.13) of the Dibranchiata"; or later: "..... The genera Orthoceras, Cyrtoceras and Gyroceras show at least the way in which the Nautilus shell can be derived from straight shells.” This illustrates the difference between an idealistic and phylogenetic approach. 71 77 78 MAIN SECTION Chapter 1 CLASS CEPHALOPODA Cuvier, 1798 Contents: a. Diagnosis. — b. Typical structure of the adult animal.— c. Typical postembryonic development. a. DIAGNOSIS Recent and, especially, fossil molluscs, as far as is known, with one or more series of arms around the mouth, a bipartite jaw apparatus resembling an inverted parrot beak, large protruding camera eyes surrounded or enveloped by the basal parts of some arms, and a funnel apparatus which narrows the mantle slit and consists of two large lateral valves (funnel pockets) which open inward, and a ventral funnel composed of two adjacent or fused muscular lobes and serving for the expulsion of water from the mantle cavity. Large, yolk-rich, symmetrically cleaved eggs which develop into young animals without a trochophorelike larval stage. Usually witha calcareous shell, chambered by regular septa, apparently pierced by a Ssiphuncle; shell straight-conical, curved or spiral. b. TYPICAL STRUCTURE OF THE ADULT ANIMAL (Including Our Concept of the Structure of the Ancestral Form) An attempt was made in the introduction to clarify the general principles of systematic morphology and to determine the special appli- cation of these principles to our subject. The following discussion will de- velop this foundation further, although it makes little difference for our purpose if we proceed according to the principles of idealistic morphology or from the standpoint of phylogenetics. At any rate, an attempt will be made to describe the form (ideal, hypothetical or real) from which Nautilus, the fossil Tetrabranchiata and the typical Dibranchiata could be derived in the simplest and most natural way. This has to be done step by step. First, the shell, the most character- istic element of a mollusc, will be studied and its general form character- ized. Then an attempt will be made to correlate it with the soft body, i.e. to correct the known structure of Nautilus so that it would fit the typical shell form and to provide the simplest possible connection between the typical structure of Dibranchiata and the general type of molluscs. With Te 19 80 its specific (involute) shell form, Nautilus appears as a much more modified and specialized cephalopod than the type of Dibranchiata established by comparative embryology. However,a number of atypical characters of Nautilus have to be eliminated to obtain the basic form. It should be remembered that the genus Nautilus can be regarded as morphologically primary in view of its great geological age (Lower Silurian), systematic position (Nautiloidea) and great resemblance to embryonic characters of Dibranchiata. This justifies special emphasis on the characteristics of Nautilus, although there are no convincing reasons for it on the principle of typical correlation (see p.23). Moreover, Nautilus is the only sur- viving cephalopod with a well developed, completely divided external shell, and the only tetrabranchiate of which the soft parts are known. It therefore provides the basis for the construction of the general type of Cephalopoda. To save space, not every detail leading to the determination of the type will be discussed. This would involve endless repetition and a complicated network of factual and logical considerations. Such proofs will therefore be given only in the most important parts; in the remaining cases, the explanation, if not self-evident, will be merely suggested. The same deduc- tive method of presentation will be followed in the next general chapters (see p.41). Thus, the proof of the type valid for a higher taxonomic group will be largely based on what is assumed for the lower groups down to species level. Because of the concise nature of the material, the reader should be able to arrive at pertinent conclusions on the basis of the general considerations if he follows the general train of thought. A part of the following discussion will have to include a repetition of the description of Nautilus. In many cases the concepts are based on embryological data, and there are therefore frequent references to Volume II where some data on the external development of Dibranchiata are described in comparison with Nautilus as indications of the primary structure of cephalopods. On the basis of paleontology, comparative anatomy and embryology, the ancestral form of Cephalopoda can only be conceived as a reconstruction of the little known genus Orthoceras, and it is therefore named Protor- thoceras. This excludes temporarily the extinct Endoceratidae, such as Endoceras, Piloceras and related forms with their wide marginal siphuncle which occupies the whole embryonic shell. There is not only the possibility but even the probability that these genera showed still earlier a morphologically primary condition from which the genus Orthoceras may have devolved. Their position is still uncertain (see p.75). At any rate, all other recent and fossil cephalopods, especially Nautilus, on the one hand, and the Dibranchiata on the other, have to be traced back to a form close toOrthoceras; the so-called "outsiders" cannot be considered here. After this qualification, it can easily be shown (p. 74) that the typical cephalopod shell is Orthoceras-like. Such shells are known from the Lower Cambrian (Volborthella), while curved and then Nautilus-like shells appear only in the Silurian. It is now established, with respect to the recent Nautilus, that, in contrast to the true ammonites or typical Gastropoda with similar shells (cf. Naef, 1911, p. 94), the juvenile or embryo- nic shells are at first not tightly coiled but merely horn-shaped like Cyr- toceras. Indirect evidence for this is found in the nucleus of the adult shell 73 (Figure 8) since very young Nautilus specimens have not yet been ob- tained. Thus,a less marked curvature has to be regarded as typical for nautiloid shells. Also the earliest Dibranchiata (Belemnoidea) always have straight phragmocones which may be curved in the young, but only in the ventral direction, i.e. in the opposite direction to the curvature of the Nautiloidea. The ancestral form can therefore only be conceived as straight. A quite clear picture of the soft parts of the ancestral form can also be obtained (Figure 19). (79) thts a RS