252 M. Edmond Perrier on (he 
XXIX. — 7^6 Origin of the Vertebrates. 
By Edmond Perrier*. 
Since the researches of Kowalevsky (1866) the Vertebrates 
have been successively derived by authors from indeterminate 
animals (Scolecida), which were considered to have given 
rise at the same time to the Tunicates (Hasckel, 1866) ; from 
the AnneHd Worms (Semper, Balfour, 1874; Dohrn, 1875; 
E. Perrier, 1881 ; Leydwich Minot, 1897) ; from the primary 
!Merostomata (Albert Gaudry, 1883) ; from Balanoglosms 
(Bateson, 1884) ; from the Nemertines (Hubrecht, 1887) ; 
from the Arachnids (Patten, 1891) ; from the Crustaceans 
(Gaskell, 1891) ; from an animal allied to the Appendicu- 
larid« (Brooks, 1893 _; Willey, 1894). 
Differences of opinion such as these evidently imply either 
that the principles of zoology are still badly defined, or else 
that they are too frequently lost sight of, or else again that 
sufficient attention has not been paid to determining the 
nature of the characters of Vertebrates, the explanation of 
■which had to be demanded from ancestral forms. We hope 
to show in the present paper that a rigorous application of 
undisputed principles leads to a unique solution of the 
problem, and that this solution is in every respect satisfactory. 
The following essential characters are exhibited by every 
Vertebrate : — 
(1) The body is bilaterally symmetrical and metamerically 
segmented throughout its entire length ; (2) an important 
extent of its external or internal surfaces, especially those 
devoted to the respiratory function, is clothed with vibratile 
cilia ; (3) during tlie embryonic period, at the very least, the 
anterior region of the alimentary canal always communicates 
with the exterior by means of lateral clefts ; (4) the circu- 
latory apparatus is closed, and exhibits a heart situated below 
the alimentary canal ; (5) the secretory apparatus is consti- 
tuted by a system of ducts which are repeated, in the embryo, 
throughout the whole length of the body, and provide the genital 
apparatus with its excretory ducts ; (6) above the digestive 
tract there extends, in the embryo, a solid cord of cells, the 
notochord^ around which are formed the vertebra? of the adult 
animal; (7) above thenotochord lies the central nervous system, 
all on the same side of the alimentary canal, deprived of an 
cesophageal ring, and exhibiting a considerable bulk ; (8) with 
* Translated by £. E. Au?ten from the ' Comptes Rendus,' t. cxxvi. 
no. 21 (May 23. 18PS), pp. U79-1486. 
Origin of the Vertehrates. 253 
reference to tlie outer world, the heart and the longitudinal 
nervous axis occupy contrary positions in the Vertebrates and 
the segmented Invertebrates, so that if we apply the term 
ventral to the side of the body turned towards the ground, 
and call the opposite side dorsal, the nervous axis is dorsal in 
the Vertebrates and ventral in the segmented animals, and 
the contractile vessels are situated on the side of the body 
opposite to the nervous system. 
Thpse are the only characters common to all Vertebrates, 
mclwd'ing AinpJiio.tus ; and since no one disputes that all these 
animals can easily be derived from the simplest forms among 
them, once these eight groups of characters are explained it 
must be admitted that the theory of the Vertebrate is com- 
plete ; conversely, every system of genealogy wliich should 
fail to explain these eight groups of characters must be 
rejected. 
(1) Metanterism of the Vertebrate Body. — Whatever be the 
cause of metamerism — a cause that we shall examine in a 
subsequent communication — the formation of the segments in 
all the metamerically segmented animals is a precocious 
developmental phenomenon, the mechanism even by which 
the body is constituted ; by this process many organisms 
are formed the segments of which may afterwards become 
obliterated, but an organism not formed by this process once 
constituted never divides again into well-defined segments. 
The law of patrogony (repetition of the genealogy by the 
embryogeny) — a fundamental law universally accepted — is 
therefore opposed to the possibility of attributing to the Verte- 
brates an ancestor whose body was not clearly segmented in 
the adult state, or at the very least multisegmental during 
the embryonic period. This at once excludes the Nemertines, 
Balanoglossus, and the Appendicularidfe, and leaves only the 
Arthropods and the Annelid Worms, In trutii, in the 
embryos of these animals the septa between the segments are 
complete, while in those of the Vertebrates they are confined 
to the dorsal half of the body ; but the embryogeny of Amphi- 
oxus shows that this is a result of tachygeneaia or emhryogenic 
acceleration. The segmentation of the embryos of Amphioxus 
is at first complete (Hatschek), and consequently identical 
with that of the Annelid Worms ; afterwards the ventral 
portion of tiie septa is absorbed again ; this condition is 
realized at once in the Vertebrates proper. 
(2) The Vibratile Cilia. — The entire organization of 
Arthropods is to some extent dominated by the property 
which their epithelial elements possess of forming, in their 
superficial region, a deposit of chitin, which causes this region 
254 M. Edmond Perrier on the 
to mortify, and renders it unsuited to the development of 
vibratile cilia. This pro|)erty manifests itself in these animals 
almost at the very beginning of the development of the 
embryo [Nanplius) ; it has rendered necessary the ecdyses 
which have in their turn occasioned the metamorphoses ; the 
absence of the cilia has had to be compensated for by the 
formation of jointed legs, moved by striped muscles and 
liavino- the respiratory apparatus dependent upon them. 
Commencing at any rate from the precocious period which 
the A'aznj/^/s represents in their ontogeny, the evolution of the 
Arthropods has therefore taken quite a peculiar direction, and 
they have remained isolated from all animals the epithelia of 
which have continued to be wholly or partially ciliated. \f the 
law of patrogony is correct, which no one disputes, there 
cannot exist beyond the Rotifers (Scirtopoda) a transitional 
form between them and the long uninterrupted series of the 
Nephridiates which proceeds from the Rotifers to the Verte- 
brates inclusively. Tiiis excludes them from tlie lineage of 
the Vertebrates, the apparent resemblances between which 
and the Merostomata, the Arachnids, or the Crustaceans are 
only cases of convergence. In particular the protective 
shields of the placoid tishes are actual bones formed in the 
derm, and not a simple epidermic covering like the pieces of 
the Arthropod carapace. 
(3) Branchial Clefts. — The presence of lateral branchial 
clefts in Balanoglossus and in the Appendicularidas has been 
one of the great arguments that have been invoked in favour 
of their relationship with the Vertebrates, This relationship 
is undeniable in the case of the Appendicularidte, but the 
absence of embryonic segmentation in these animals proves 
(law of patrogony), as we have already pointed out, that they 
are not ancestral Vertebroids, but degraded Vertebroids ; the 
same remark would apply to Balanoglossus , in case we were 
to admit as real the resemblances which people have striven 
to find between it and the Vertebrates, apart from its branchial 
clefts. But this latter resemblance is itself open to suspicion. 
The branchial clefts of the Vertebrates and the primary and 
secondary branchial clefts of Amphioxus are repeated, in fact, 
exactly like the metaraeres; although the embryo of Balano- 
glossus exhibits manifest traces of metameric segmentation, 
there is no relation between the metameres and the branchial 
clefts. This would be explained, for once in a way, as is 
shown by the discord which iinally supervenes in the case of 
Aviphiocuus itself, under the supposition of a degeneration on 
the part of Balanoglossus^ but not on the hypothesis which 
makes it an anccatral form. The absence of branchial clelts 
Origin of the Vertebrates. 255 
in the Annelid Worms is an objection that may be lodged 
against the annelidan theory of the Vertebrates. It is easily 
disposed of. The branchial clefts are, in fact, but adven- 
titious orifices of lateral diverticula of the alimentary canals. 
Now many Annelid Worms exhibit such diverticula ; they 
are also produced in the Turbellaria and the Nudibranch 
Mollusks, which are nephridiates, and in this case these diver- 
ticula may open to the exterior {Yungia, Cgcloporus, 
Balanoglossus itself, were it tlie ancestor of the Vertebrates, 
would come to be intercalated between them and the Annelid 
Worms, and would bear witness to the possibility of the 
appearance of pharyngeal clefts in these latter. The pheno- 
mena of budding in various CTEXODRiLiDiE, SyllidtE, and 
Serpulix^, as well as in the Natdomorpha, also demon- 
strate the possibility of the appearance of adventitious orifices 
to the digestive canal in the case of the Annelid Worms 
properly so-called. 
(4) Circulatory Apparatus. — Among the ancestors attri- 
buted to the Vertebrates, the Annelid Worms, the Nemer- 
tines, and BaJannglossus have, like the Vertebrates them- 
Kelves, a closed circulatory apparatus ; on the other hand, no 
Arthropod is known the circulatory apparatus of which is not 
in part lacunar. On this head the Arthropods are therefore 
excluded for the second time. The circulatory apparatus of 
the Nemertines has no differentiated propulsive centre ; that 
which is called the heart in Balanoglossus is an organ situated 
on the same side of tlie alimentary canal as the nervous axis, 
and cannot consequently be homologous with the heart of 
the Vertebrates ; the Annelid Worms tiierefore alone exhibit 
the typical relations of the circulatory centres, alimentary 
canal, and nervous axis, which are observed in the case of the 
Vertebrates. 
(5) Nephridial Apj^arafus. — The urinary apparatus of the 
true Arthropods* is constructed after quite a different type from 
that of the uephridial apparatus of tiie Annelid Worms, the 
Nemertines, Anqihioxus, and the Vertebrates; this apparatus 
appears to be wanting in Balanoglossus, \y
c\\ thus furnishes 
a new motive for its exclusion ; in the case of the Annelid 
Worms and in Amphioxus it is formed of independent ciliated 
tubes, opening into the general cavity on the one side and to 
the exterior on the otiier, and recurring regularly either in 
almost all the segments of the body or only in a certain 
* Among these Peripatus cannot be included with certainty, in spite 
of the attempts that have recently been made to connect it with them. 
256 M. Edmond Perrier on the 
number of them, the anterior segments, for instance (Amphi- 
CTENID.E, Ampharetid^, Terebellid.e, Amphioxus). The 
nephridia of the Nemertines are also contined to the anterior 
region of the body, and eventually open into a collecting duct 
on each side ; herein they seem to present a farther resem- 
blance to those of tlie Vertebrates : but analogous conditions 
exist in the case of the Annelid Worms [Loimia medusa, 
Lanice conchyhga) ; in certain Lumbricidfe [Octochcetus, Peri- 
chceta) and some Hirudineans [PontohdtUa) the nephridia 
form throughout the body but one and the same system of 
reticulated tubes. Moreover, in tiie Annelid Worms the 
nephridia almost always place themselves more or less at the 
service of the genital apparatus, as in the case of the Verte- 
brates. In the Earthworms (LumbricIMORPHA) there even 
occurs a doubling of the renal apparatus and of the excretory 
apparatus of the genital glands, the analogies between which 
and that whicli is observed in the case of the Vertebrates 
(Wolffian and Miillerian ducts) t set forth as early as 1881 ^. 
The advantage therefore rests with the Annelid Worms, of 
which the Nemertines are in all probability nothing but 
highly modified forms. The Annelid Worms being hence- 
forth alone concerned, it remains for us to explain how it has 
been possible for them to be the point of departure for the 
organic conditions which are peculiar to the Vertebrates. 
(6) 2iotochord. — If it has been possible to compare for one 
moment the endodermic diverticulum of the proboscis of 
Balanoglossus, or even the proboscis of the Nemertines with 
tlie notochord of the embryos of Vertebrates and of the Tuni- 
cates, it is in default of having rigorously defined the con- 
ditions which have brought about the formation of the 
notochord and the necessary relations which result from this 
formation. The embryogeny of AmpJn'oxiis, like that of the 
Tunicates, shows in fact that tlie notochord is originally not 
a cord ot cells, but the entire region of the endoderm included 
between the rudiment of the nervous system and the two 
rudiments of the mesoderm. These three rudiments are 
three regions of active development, which are only able to 
derive the reserve of nutriment which is necessary for them 
from the endodermic area comprised between them. As a 
result of this the elements belonging to the area thus circum- 
scribed empty themselves, become vacuolate and to a certain 
extent mortified ; the rest of the endoderm, on the contrary, 
continues to develop, glides in consequence underneath the 
inert area and eliminates it from the endoderm, and it is this 
* ' Les Colonies aiiimales,' pp. 677 and (38 1, 
Oriyin of the Vertehraies. 257 
eliminated area, tlie origin of which is quite clear, wliich, by 
the grouping of its elements into a cord, constitutes the noto- 
ciiord. The cord thus constituted has nothing except the 
mortification of its cells in common with the endodermic 
diverticulum wdiich penetrates into the proboscis of Bilaao- 
glossus and which has its analogue in the case of Cephalo- 
discus. These structures can be no more likened to a noto- 
chordal rudiment than could the median diverticulum, entirely 
analogous this time to that in Balanoglossus^ which is thrown 
out in front by the main intestine of the majority of the poly- 
clad and triclad Turbellarians. On the contrary, it will 
easily be understood how it has been possible for a notachord 
to be formed in an Annelid Worm, of whicli the nervous 
system, developed to an exceptional degree, and the nieso- 
dermic rudiments have been affected by tachygenesis. Now 
the formation of the nervous system by an invagination of 
the ectoderm and that of the mesoderm by two evaginations 
of the endoderm are evident proofs of the intervention of 
tachygenesis *. The considerable proportions assumed by 
the nervous system furnish, moreover, the explanation of the 
two latter distinctive characteristics of the Vertebrates. 
(7) Disappearance of the (Esophageal Ring. — The excep- 
tional volume assumed by the nervous system in all Verte- 
brates excludes from their genealogy Balanoglossus, in which, 
on the contrary, the axial cord is reduced to a very short 
rudiment which does not exceed the collar in length. The 
proportions to which the nervous axis attains necessarily 
induce its precocious development, its formation by rapid 
processes. It originates, in fact, at the expense of a fairly 
large area of the ectoderm, which buries itself beneath the 
neighbouring regions at a period well in advance of that at 
which the mouth is constituted. The nervous system beino- 
outlined before the mouth has no longer to concern itself witti 
the existence of the latter, as it does in the case of the 
Annelid Worms, by developing around it. There is no 
longer a raison d'etre for the oesophageal ring f y it disap- 
pears, and the brain proceeding to its completion on the 
median neural line stands in the way of the formation of the 
mouth upon this line. It is the cause of the reversal of the 
attitude of the Vertebrates, as already pointed out by 
Geoffroy Saint-Hilaire. 
(b) Mever&al of Attitude in the Vertebrate. — The mere com- 
parison of a section of the embryo of a dog-fish and of a 
* See ' Comptes Rendus,' December 1896 : Report on the Competition 
for the Serres Prize, by the author. 
t E. Perrier, ' Les Colonies animales,' p. 695 (1881). 
258 On the Orlyln of the Vertebrates. 
section of an Annelid \^'orm suffices to demonstrate, as was 
proved by Semper, after GeofFroy Saint-Hilaire, that the in- 
verse arrangement of the organs in the two branches is easily 
explained by a simple change of attitude. If we consider the 
embryogeny of Amphioxus, such as it has been described by 
Hatschek and, most recently, by Willey, keeping before us the 
considerations of which a resume has been given above, all 
the apparently inexplicable peculiarities in the development 
of this animal not only become cleared up of themselves, but 
show the way in which the inversion of the higher Verte- 
brates has come to pass. The mouth of Amphioxus does not 
pass indeed all at once to the median line, opposite to the 
line of the nervous axis ; it is formed at the nearest possible 
point to its primitive situation, on the left side of the body. 
The mouth having become lateral, the animal cannot feed 
except on condition that it lies on the left side of the body, 
which becomes its habitual attitude. According to Lamarck's 
principle (use a7id disuse of organs), of which the entire 
liistory of the Invertebrates is also a striking confirmation, it 
results from this new attitude that the organs of sense of the 
left half of the body, being more in touch with the nutriment 
supplied by the food, alone develop : the young animal 
exhibits but one olfactory pit and one gustatory organ 
(Hatschek's organ), both situated on the left. In consequence 
of this same attitude all the branchial clefts on the left side 
are covered up, and it is rendered impossible for them to 
perform their normal function. The animal is therefore 
induced to twist the region of its body w^hich corresponds to 
them, so as to bring all its branchial clefts on to the right 
side. This is no mere hypothesis ; this torsion is exhibited 
by the young Amphioxus in the course of its development 
(Willey), without any present physiological necessity being 
able to explain it ; its two series of branchial clefts, the 
projecting folds between which they ai'e enclosed, which are 
the first traces of the subbranchial cavity of the adult animal, 
are formed on the right side of the body exclusively and mark 
out the curves along which the torsion has been accomplished. 
The application of the law of patrogony necessitates the 
conclusion that the ancestors of Amphioxus passed through a 
period at which they lived lying on the left side, with the 
mouth in contact with the ground, and when they were 
obliged, in order to ensure the regularity of their respiration, 
to twist the branchial region of the bod}'. This permanent 
attitude has produced, in consequence of the constant contrac- 
tion of certain muscles and the relaxation of those opposed to 
them, an asymmetry of the body which has been transmitted 
On new Coleoptera from East Africa. 2o9 
by heredity {Lamarck's principle), and this phase in tlie 
history of Amphioxus is still found inscribed upon its 
embryogenic development. Later on the tinimal adopted the 
habit of burying its posterior extremity in the sand; it found 
itself subjected to a life in a homogeneous medium ; the sym- 
metrical shapes, acting in harmony with a more ancient heredity 
unopposed by contrary efforts, and in accordance even with the 
conformation of the regions of the body which have escaped 
the torsion, caused the mouth, by means of successive distor- 
tions, to take its place in the plan of symmetry, and, since it 
was unable to regain its position in the dorsal median line, it 
passed over to the ventral side. Thus was brought about the 
new attitude and the return to a perfect symmetry on the part 
of the Vertebrates descending from Amphioxus or from 
analogous Vertebroids *. 
XXX. — Descriptions of new Coleoptera from East Africa. 
By C. O. Waterhouse, F.E.S. 
The following interestuig new Coleoptera have been lent me 
for description. 'l-^l»ey form part of a valuable collection 
made by Mr. Hollis in Usambara in 189o and 1894. 
* Amphioxus iu the adult state has not yet recovered this symmetry, 
as is shown by the presence on the left side of the body of a single olfac- 
tory pit, the continuity of the right half of the oral hood with the cephalic 
expansion of the fin, the continuity of the right metapleure alone with 
the ventral fin, the overlapping of the right and left muscle-segments, 
and, finally, the abortion of the genital organs of the left side la 
Amj}hiodU8 cultellus, from Torres Straits, and A. lucayanus, from the 
Bahamas, for which have been founded the genera Epiyonichthys and 
Asymmetron. 
The inductions by means of which we have interpreted the phenomena 
of torsion which are proved to take place in the development of Amphi- 
oxus, and the deductions which we have drawn from these phenomena 
with reference to the history of the reversal of the Vertebrates, are but 
strictly scientific applications of general laws, recognized by all in prin- 
ciple, but too often forgotten in each particular case. We observe, more- 
over, phenomena of torsion, which are analogous to the above and are 
explained by the principle of Lamarck, likewise in the development of 
many fixed animals (Cirrbipedes, Bryozoa, Echinoderms, Tunicates), in 
the case of the Gastropod moUusks and the Pleuronectid fishes. The rule 
■which governs all these phenomena, and which may be termed the rule 
of the Jixation of attitudes, may be expressed thus: — 
When, in passing from 07ie mode of life to another, an animal is led, in 
order to ensure that its organs shall perform their functions properly, to 
assume habitually a definite attitude, this attittide is capable of becoming 
Axed and hereditarily transmitted.