The “titillator” (Brunner von Wattenwyl) is a solid curved hook with a hollow base. Besides these, are several odd-shaped, unsymmetrical pieces (fig. 102, C, D, E, F), moved by special muscles. A pair of styles (see figs. 32–3 and 103) project from the hinder edge of the 9th sternum. These paired and unpaired appendages are believed to open the genital pouch of the female, but we do not understand their action in detail.173

Brehm observes that the male reproductive organs of the Cockroach are most nearly paralleled by those of the Mantidæ. A free penis occurs in all Orthoptera, except Acridiidæ and Phasmidæ.

The male organs of the House Cricket will be found much easier to understand than those of the Cockroach. The testes are of irregular, oval figure, the vasa deferentia very long, tortuous, and enlarged towards the middle of their length. The vesiculæ seminales bear many utriculi majores et breviores. The penis is of simple form, and dilated at the end. The titillator is broad, but produced into a slender prong, which projects beyond the penis. A pair of subanal styles is found, but the unpaired hooklets are wanting or very inconspicuous.

Very little is known about the act of copulation among Cockroaches, and the opportunities of observation are few. The following account is given by Cornelius (loc. cit., p. 22):—

CHAPTER X.

The Embryonic Development of the Cockroach. 174

By JOSEPH NUSBAUM, Magister of Zoology, Warsaw.

The development of the Cockroach is by no means an easy study. It costs some pains to find an accessible place in which the females regularly lay their eggs, and the opaque capsule renders it hard to tell in what stage of growth the contained embryos will be found. Accordingly, though the development of the Cockroach has lately attracted some observers, the inexperienced embryologist will find it more profitable to examine the eggs of Bees, of Aphides, or of such Diptera as lay their eggs in water.

The Cockroach is developed, like most animals, from fertilised eggs.175 The eggs of various animals differ much in size and form, but always contain a formative plasma or egg-protoplasm, a germinal vesicle (nucleus), and a germinal spot (nucleolus). Besides these essential parts, eggs also always contain a greater or less quantity of food-yolk, which serves for the supply of the developing embryo. The quantity of this yolk may be small, and its granules are then uniformly dispersed through the egg-protoplasm; or very considerable, in which case the protoplasm and yolk become more or less sharply defined. Eggs of the first kind are known as holoblastic, those of the second kind as meroblastic, names suggested by the complete or partial segmentation which these kinds of eggs respectively undergo. When the food-yolk is very abundant it does not at first (and in some cases does not at any time) exhibit the phenomena of growth, such as cell-division. If, on the other hand, the yolk is scanty and evenly dispersed through the egg-protoplasm, the segmentation proceeds regularly and completely. The eggs of Arthropoda, including those of the Cockroach, are meroblastic.

The eggs of the Cockroach (P. orientalis) are enclosed (see p. 23) sixteen together in stout capsules of horny consistence. They are adapted to the form of the capsule, laterally compressed, convex on the outer, and concave on the inner side. The ventral surface of the embryo lies towards the inner, concave surface of the egg. Each egg is provided with a very thin brownish shell (chorion), whose surface is ornamented with small six-sided projections. In young eggs, still enclosed within the ovary, the nucleus (germinal vesicle) and nucleolus (germinal spot) can be plainly seen, but by the time they are ready for deposition within the capsule, so large a quantity of food-yolk, at first finely—afterwards coarsely—granular, accumulates within them, that the germinal vesicle and spot cease to be visible.

Since the yolk of the newly-laid egg of the Cockroach is of a consistence extremely unfavourable to hardening and microscopic investigation, I have not been able to obtain transverse sections of the germinal vesicle, nor to study the mode of its division (segmentation). If, however, we may judge from what other observers have found in the eggs of Insects more suitable for investigation than those of the Cockroach, we shall be led to conclude that a germinal vesicle, with a germinal spot surrounded by a thin layer of protoplasm, lies within the nutritive yolk of the Cockroach egg. From this protoplasm all the cells of the embryo are derived.

The germinal vesicle, together with the surrounding protoplasm, undergoes a process of division or segmentation. Some of the cells thus formed travel towards the surface of the egg to form a thin layer of flattened cells investing the yolk, the so-called blastoderm, while others remain scattered through the yolk, and constitute the yolk-cells (fig. 107).

On the future ventral side of the embryo (and therefore on the concave surface of the egg) the cells of the blastoderm become columnar, and here is formed the so-called ventral plate, the first indication of the embryo. This is a long narrow flattened structure (fig. 104). It is wider in front where the head segment is situated; further back it becomes divided by many transverse lines into the primitive segments. The total number of segments in the ventral plate of Insects is usually seventeen.176 Indications of the appendages appear very early. They give rise to an unpaired labrum, paired antennæ, mandibles, and maxillæ (two pairs). The first and second pair of maxillæ have originally, according to Patten,177 two and three branches respectively. Behind the mouth-parts are found three rudimentary legs. Upon all the abdominal segments, according to Patten, rudimentary limbs are formed; but these soon disappear, except one pair, which persists for a time in the form of a knobbed stalk; subsequently this, too, completely disappears. Three or four of the hindmost segments curve under the ventral surface of the embryo, and apparently (?) give rise to the modified segments and appendages of the extremity of the abdomen (fig. 105). The ventral plate lies at first directly beneath the egg membrane (chorion), but afterwards becomes sunk in the yolk, so that a portion of the yolk makes its way between the ventral plate and the chorion. Whilst this portion of the yolk is perfectly homogeneous, the remainder, placed internally to it, becomes coarsely granular, and encloses many roundish cavities and yolk-cells. The middle region of the body is more deeply sunk in the yolk than the two ends, and the embryo thus assumes a curved position (fig. 105).

As to the origin of the mesoblast most observers have found178 that a long groove (the germinal groove) appears in the middle line of the ventral plate (fig. 108), which bulges into the yolk, gradually detaches itself from the epiblast, and completes itself into a tube. The lumen of this tube soon becomes filled with cells, and the solid cellular mass thus formed divides into two longitudinal tracts, which lie right and left of the middle line of the ventral plate beneath the epiblast, and are known as the mesoblastic bands. In the Cockroach I was able to satisfy myself that in this Insect also, the mesoblast, in all probability, arises by the formation and closure of a similar groove of the epiblast. M (fig. 108) represents the stage in which the lumen of the groove has disappeared, and the mesoblast forms a solid cellular mass.

The origin of the hypoblast in Insects has not as yet been clearly determined. Two quite different views on this subject have found support. Some observers (Bobretsky, Graber, and others) maintain that the hypoblast originates in the yolk-cells, which form a superficial layer investing the rest of the yolk. Others (especially Kowalewsky179) believe that the process is altogether different. According to the latest observations of the eminent embryologist just named, upon the development of the Muscidæ, the germinal groove gives rise, not only to the two mesoblastic bands, but also, in its central region, to the hypoblast. This makes its appearance, however, not as a continuous layer, but as two hourglass-shaped rudiments, one at the anterior, the other at the posterior end of the ventral plate. These rudiments have their convex ends directed away from each other, while their edges are approximated and gradually meet so as to form a continuous hypoblast beneath the mesoblast. Although I have not been able completely to satisfy myself as to the mode of formation of the hypoblast in the Cockroach, I have observed stages of development which lead me to suppose that it proceeds in this Insect in a manner similar to that observed by Kowalewsky in Muscidæ. The hourglass-shaped rudiments of the hypoblast become pushed upwards by those foldings-in of the epiblast which form towards the anterior and posterior ends of the embryo, and give rise to the stomodæum and proctodæum.180

The stage of development in which the germinal groove appears, by the folding inwards of the epiblast, has been observed in many other animals, and is known as the Gastræa-stage. In all higher types (Vertebrates, the higher Worms, Arthropoda, Echinodermata) the mesoblast and hypoblast are formed in the folded-in part of the Gastræa in a manner similar to that observed in Insects.

The yolk-cells, which some observers have supposed to form the hypoblast, are believed by Kowalewsky to have no other function except that of the disintegration and solution of the yolk. I can, however, with confidence affirm that in the Cockroach these cells take part in the formation of permanent tissues (see below).

Each of the two mesoblastic bands which lie right and left of the germinal groove divides into many successive somites, and each of these becomes hollow. Every such somite consists of an inner (dorsal) one-layered and an outer (ventral) many-layered wall, the latter being in contact with the epiblast. The cavities of all the somites unite to form a common cavity, the cœlom or perivisceral space of the Cockroach. The cœlom, like the cavities in which it originates, is bounded by two layers of mesoblast—an inner, the so-called splanchnic or visceral layer, which lies on the outer side of the hypoblast, and an outer somatic or parietal layer, beneath the epiblast. There are accordingly four layers in the Cockroach-embryo—viz., (1) epiblast, from which the integument and nervous system are developed; (2) somatic layer of mesoblast, mainly converted into the muscles of the body-wall; (3) splanchnic layer of mesoblast, yielding the muscular coat of the alimentary canal; and (4) hypoblast, yielding the epithelium of the mesenteron.

Scattered yolk-cells associate themselves with the mesoblast cells, so that the constituents of the mesoblast have a two-fold origin. Fig. 109 shows that the yolk-cells are large, finely granular, and provided with many (3–6) nuclei and nucleoli. They send out many branching protoplasmic threads, which connect the different cells together, and thus form a cellular network. Certain cells separate themselves from the rest, apply themselves to the walls of the somites, and form a provisional diaphragm (fig. 110, D) consisting of a layer of flattened cells;181 other cells (fig. 109) pass into and through the walls of the somites, and reach their central cavity, where they increase in number and blend with the mesoblast cells. What finally becomes of them I cannot say; perhaps they form the fat-body.

The ventral plate occupies, as I have explained, the future ventral surface of the Insect, and here only at first both the embryonic membranes are to be met with. On the sides and above the yolk is invested by the serous envelope alone. The ventral plate, however, gradually extends upwards upon the sides of the egg, in the directions of the arrows (fig. 107), and finally closes upon the dorsal surface of the embryo, so as completely to invest the whole yolk. Every segment of the embryo shows at a certain stage numerous clusters of spherical granules, which according to Patten (loc. cit.) are composed of urates (fig. 111, S).

We shall now proceed to consider the development of the several organs of the Cockroach.

Nervous System.—Along the middle line of the whole ventral surface there is formed a somewhat deep groove-like infolding of the epiblast, bounded on either side by paired solid thickenings, which detach themselves from the epiblast (fig. 110, N) and constitute the double nervous chain. In many other Insects a median cord (from which are derived the transverse interganglionic commissures) forms along the bottom of the nervous fold. This secondary median fold is very inconspicuous and slightly developed in the Cockroach, so that the transverse commissures between the developing ganglia are mainly contributed by the cellular substance of the lateral nervous band. The brain is formed out of two epiblastic thickenings which occupy shallow depressions. The so-called inner neurilemma, which surrounds the ventral nerve-cord, is developed as follows:—Along the ventral nerve-cord, and between its lateral halves, a small solid cellular band (fig. 110, Ch) is developed out of the mesoblastic diaphragm described above. This grows round the ventral nerve-cord on all sides (fig. 112, N′), passing also inwards between the central fibrillar tract and the outer cellular layer, and thus forming the thin membrane which invests the central nervous mass (fig. 112, N″). The above-mentioned solid mesoblastic band, which exists for a very short time only, may perhaps be homologised with the chorda dorsalis of Vertebrates, and the chorda of the higher Worms, since in these types also the chorda forms a solid cellular band of meso-hypoblastic origin, lying between the nervous system and the hypoblast. The peripheral nerves arise as direct prolongations of the fibrillar substance of the nerve-cord.

Alimentary Canal.—The epithelium of the mesenteron is formed out of the hypoblast, whose cells assume a cubical form and gradually absorb the yolk. The epithelium of the stomodæum and proctodæum is derived, however, from two epiblastic involutions at the fore and hind ends of the embryo. The muscular coat of the alimentary canal is contributed by the splanchnic layer of the mesoblast. The mesenteron in an early stage of development appears as an oval sac of greenish colour (fig. 113), faintly seen through the body-wall. The cæcal tubes are extensions of the mesenteron, the Malpighian tubules of the proctodæum. The epiblastic invaginations may be recognised in all stages of growth by their chitinous lining and layer of chitinogenous cells, continuous with the similar layers in the external integument.

Tracheal System.—Tubular infoldings of the epiblast, forming at regular intervals along the sides of the embryo and projecting into the somatic mesoblast, give rise to the paired tracheal tubes, which are at first simple and distinct from one another.182

Heart.—The wall of the heart in Insects is of mesoblastic origin, and develops from paired rudiments derived from that peripheral part of each mesoblastic band which unites the somatic to the splanchnic layer. In this layer two lateral semi-cylindrical rudiments appear, which, as the mesoblastic bands meet on the dorsal surface of the embryo, are brought into contact and unite to form the heart (fig. 111). The heart is therefore hollow from the first, its cavity not being constricted off from the permanent perivisceral space enclosed by the mesoblast, but being a vestige of the primitive embryonic blastocœl, which is bounded by the epiblast, as well as by the two other embryonic layers. Such a mode of the development of the heart was observed by Bütschli in the Bee, and by Korotneff in the Mole Cricket. I am convinced, from my own observations, that the heart of the Cockroach originates in this way, though it is to be observed that, in consequence of Patten’s results,183 the question requires further investigation. According to Patten the mesoblastic layers of the embryo pulsate rhythmically long before the formation of the heart. Patten also states that the blood-corpuscles are partially derived from the wall of the heart.

Reproductive Organs.—In P. orientalis the reproductive organs are developed as follows:—The reproductive glands have a mesoblastic origin. The immature ovaries and testes take the form of elongate oval bodies, which prolong themselves backwards into a long thin thread-like cord or ligament (figs. 114, 115). These lie in the perivisceral space, between the somatic and splanchnic layers of the mesoblast, and on the sides of the abdomen. The glands divide tolerably early into chambers, which have, however, a communicating passage (figs. 114, 115). From their backward-directed prolongations arises the epithelium of the vasa deferentia and oviducts. All other parts of the reproductive ducts are developed out of tegumentary thickenings of the ventral surface in the last abdominal segment, and the last but one. These thickenings are at first paired,184 but afterwards blend to form single organs (fig. 118). Within the tegumentary thickenings just described, there appear in the male Cockroach two anterior closed cavities which unite to form the single cavity of the permanent mushroom-shaped body (vesicula seminalis). A posterior cavity becomes specialised as the ductus ejaculatorius, while the hindmost part of the thickening, which is at first double, afterwards by coalescence single, forms the penis (figs. 117, 118). The accessory reproductive glands have also a tegumentary origin. In the female Cockroach the chitinogenous epithelium of the integument gives rise to the uterus, vagina,185 and accessory glands, the muscular and connective tissue layers of the sexual apparatus being formed out of loose mesoblastic cells.186

Joseph Nusbaum.

Animal Metamorphoses.

To investigate the causes of metamorphosis, let us select from the same sub-kingdom two animals as unlike as possible with respect to the amount of post-embryonic change to which they are subject. We can find no better examples than Amphioxus and the Chick.

The newly-hatched Amphioxus is a small, two-layered, hollow sac, which moves through the sea by the play of cilia which project everywhere from its outer surface. It is a Gastræa, a little simpler than the Hydra, and far simpler than a Jelly-fish. As yet it possesses no nervous system, heart, respiratory organs, or skeleton. The most expert zoologist, ignorant of its life-history, could not determine its zoological position. He would most likely guess that it would turn either into a polyp or a worm.

The Chick, on the other hand, at the tenth day of incubation, is already a Bird, with feathers, wings, and beak. When it chips the shell it is a young fowl. It has the skull, the skeleton, the toes, and the bill characteristic of its kind, and no child would hesitate to call it a young Bird.

Amphioxus is, therefore, a Vertebrate (if for shortness we may so name a creature without vertebræ, brain, or skull), which develops with metamorphosis, being at first altogether unlike its parent. The Chick is a Vertebrate which develops directly, without metamorphosis. Let us now ask what other peculiarities go with this difference in mode of development.

Amphioxus produces many small eggs ( 1/10 mm. in diameter) without distinct yolk, and consequently segmenting regularly. The adult is of small size (2 to 3 in. long), far beneath the Chick in zoological rank, and of marine habitat.

The Fowl lays one egg at once, which is of enormous size and provided with abundant yolk, hence undergoing partial segmentation. The Fowl is much bigger than Amphioxus, much higher in the animal scale, and of terrestrial habitat.

Which of the peculiarities thus associated governs the rest? Is it the number or size of the eggs? Or the size, zoological rank, or habitat of the adult? The question cannot be answered without a wider collection of examples. Let us run over the great divisions of the Animal Kingdom, and collect all the facts which seem to be significant. We may omit the Protozoa, which never develop multicellular tissues, and in which segmentation and all subsequent development are therefore absent.

Porifera (Sponges).—Nearly all marine and undergoing metamorphosis, the larva being wholly or partially ciliated.

Cœlenterates undergo metamorphosis, the immediate product of the ovum being nearly always a planula, or two-layered hollow sac, usually devoid of a mouth, and moving about by external cilia. In many Cœlenterates the complicated process of development known as Alternation of Generations occurs. The sedentary Anemones pass through a planula stage, but within the body of the parent. Among the few Cœlenterates which have no free planula stage is the one truly fluviatile genus—Hydra.

Worms are remarkable for the difference between closely allied forms with respect to the presence or absence of metamorphosis. The non-parasitic freshwater and terrestrial Worms, however (e.g., Earthworms, Leeches, all freshwater Dendrocœla, and Rhabdocœla), do not undergo metamorphosis. In the parasitic forms complicated metamorphosis is common, and may be explained by the extraordinary difficulties often encountered in gaining access to the body of a new host.

All Polyzoa are aquatic (fluviatile or marine), and all produce ciliated embryos, unlike the parent.

Brachiopoda are all marine, and produce ciliated embryos.

Echinoderms usually undergo striking metamorphosis, but certain viviparous or marsupial forms develop directly. There are no fluviatile or terrestrial Echinoderms.

Lamellibranchiate Mollusca have peculiar locomotive larvæ, provided with a ring of cilia, and usually with a long vibratile lash. These temporary organs are reduced or suppressed in the freshwater forms. There are no terrestrial Lamellibranchs.

Snails have also a temporary ciliated band, but in the freshwater species it is slightly developed (Limnæus), and it is totally wanting in the terrestrial Helicidæ.

Cephalopoda, which are all marine, have no ciliated band, and the post-embryonic changes do not amount to metamorphosis. There is usually a much larger yolk-sac than in other Mollusca.

Crustacea usually pass through well-marked phases. Peneus presents five stages of growth (including the adult), the earlier being common to many lower Crustacea. The Crab passes through three, beginning with the third of Peneus; the Lobster through two; while the freshwater Crayfish, when hatched, is already in the fifth and last.

Fishes seldom undergo any post-embryonic change amounting to metamorphosis. Amphioxus (if Amphioxus be indeed a fish) is the only well-marked case.

Amphibia develop without conspicuous metamorphosis, except in the case of the Frogs and Toads (Anura), which begin life as aquatic, tailed, gill-bearing, and footless tadpoles.

Reptiles, Birds, and Mammals do not undergo transformation.

This survey, hasty as it necessarily is, shows that habitat is a material circumstance. Larval stages are apt to be suppressed in fluviatile and terrestrial forms. Further, it would seem that zoological rank is not without influence. Metamorphosis is absent in Cephalopoda, the highest class of Mollusca, and in all but the lowest Vertebrates, while it is almost universal in Cœlenterates, Echinoderms, and Lamellibranchs.

It has often been remarked that the quantity of food-yolk indicates the course of development. If a large store of food has been laid up for the young animal, it can continue its growth without any effort of its own, and it leaves the egg well equipped for the battle of life. Where there is little or no yolk, the embryo is turned out in an ill-furnished condition to seek its own food. This early liberation implies metamorphosis, for the small and feeble larva must make use of temporary organs. Some very simple locomotive appendages are almost universally needed, to enable it to get away from the place of its birth, which is usually stocked with as much life as it can support.

Some animals, therefore, are like well-to-do people, who can provide their children with food, clothes, schooling, and pocket-money. Their fortunate offspring grow at ease, and are not driven to premature exercise of their limbs or wits. Others are like starving families, which are forced to send their children to sell matches or newspapers in the streets. It is a question of the amount of capital or accumulated food which is at command.

The connection between zoological rank and the absence of metamorphosis is also explained by what we see among men. High zoological position ordinarily implies strength or intelligence, and the strong and knowing can do better for their offspring than the puny and sluggish. It does not cost a Shark or a quadruped too much to hatch its young in its own body, while Spiders and Earwigs,187 which are among the highest Invertebrates, defend their progeny, as do Mammals and Birds, the highest Vertebrates.

But what has all this to do with habitat? Are fluviatile and terrestrial animals, as a rule, better off than marine animals? Possibly they are. In the confined and isolated fresh waters at least, the struggle for existence is undoubtedly less severe than in the waters of the sea. This is shown by the slow rate of change in freshwater types. Many of our genera of land and freshwater shells date back at least as far as to Purbeck and Wealden times, while our common pond-mussel is represented in the Coal Measures. The comparative security of fresh waters is probably the reason why so many marine fishes enter rivers to spawn.

More important, and less open to question, is the direct action of the sphere of life. The cheap method of turning the embryo out to shift for itself can seldom be practised with success on land. But in water floating is easy, and swimming not difficult. A very slightly-built larva can move about by means of cilia, and a whole brood can disperse far and wide in search of food, while still in a mere planula condition—hollow sacs, without mouth, nerves, or sense-organs. Afterwards the little locomotive larva settles down, opens a mouth, and begins to feed. Nearly the whole of its development is carried on at its own charge.

The extra risks to which marine animals are exposed also tell in favour of transformation, for they are met by an increase in the number of ova. Marine species commonly lay more eggs than freshwater animals of like habits. The Cod is said to produce nine million eggs; the Salmon from twenty to thirty thousand; the Stickleback only about one hundred, which are guarded during hatching by the male. The Siluroid fish, Arius, lays a very few eggs, as big as small cherries, which the male carries about in his mouth.

Without laying stress upon such figures as these, which cannot be impartially selected, we can safely affirm that marine forms are commonly far more prolific than their freshwater allies. But high numbers increase the difficulty of providing yolk for each, and thus tend to early exclusion, and subsequent transformation. We may rationally connect marine habitat with small eggs, poorly supplied with yolk, segmenting regularly, and producing larva which develop with metamorphosis.

In fresh waters dispersal can seldom be very effective. The area is usually small, and communicates with other freshwater basins only through the sea. Migration to a considerable distance is usually impossible, and migration to a trifling distance use less. Moreover, competition is not too severe to prevent some accumulation of food by the parent on behalf of the family.

On land the conditions are still less favourable to larval transformation. Very early migration is altogether impossible. Any kind of locomotion by land implies muscles of complicated arrangement, and, as a rule, there must be some sort of skeleton to support the weight of the body. The larva, if turned out in a Gastræa condition would simply perish without a struggle.188 Nor is great precocity needful. The terrestrial animal is commonly of complicated structure, active, and well furnished with means of information. It can lay-by for its offspring, and nourish them within its own body, or at least by food stored up in the egg.

The influence of habitat upon development may be recapitulated as follows:—

Marine Habitat.—Eggs many. Yolk small. Segmentation often regular. Young hatched early. Development with metamorphosis. [The most conspicuous exceptions are Cephalopoda and marine Vertebrata.]

Fluviatile Habitat.—Eggs fewer. Yolk larger. Segmentation often unequal. Young hatched later. Development direct, or with late metamorphosis only. [The most obvious exceptions are Frogs and Toads, which develop with metamorphosis.]

Terrestrial Habitat.—Eggs few. Yolk large [except where the young are supplied by maternal blood]. Segmentation often partial. Young hatched late. Development without metamorphosis. [An exception is found in Insects, which usually exhibit conspicuous metamorphosis, though the yolk is large, and the type of segmentation partial or unequal.]

Let us now take up the exceptions, and see whether these are capable of satisfactory explanation.

1.—Cephalopoda and marine Vertebrates, unlike other inhabitants of the sea, develop without metamorphosis. But these are large animals of relatively high intelligence, well able to feed and protect their young until development is completely accomplished.

2.—Frogs and Toads, unlike other fluviatile animals, develop with metamorphosis. The last and most conspicuous change, however, from the gill-bearing and tailed tadpole to the air-breathing and tailless frog, hardly belongs to the ordinary period of embryonic development. When the tadpole has four limbs and a long tail it has already reached the point at which the more primitive Amphibia (Menopoma, Proteus, &c.) become sexually mature. The loss of the tail, the lengthening of the hind limbs, and the complete adaptation to pulmonary respiration, relate to the mode of dispersal of the adult. Cut off from early dispersal by the isolation of their breeding-places and the difficulty of land migration, Frogs migrate from pool to pool as full-grown animals. The eggs are thus laid in new sites, and very small basins—ditches and pools which dry up in summer—can be used for spawning. To this peculiar facility in finding new spawning grounds the Anura no doubt owe their success in life, of which the vast number of nearly-allied species furnishes an incontrovertible proof. But the adaptation to terrestrial locomotion necessarily comes late in life, after the normal and primitive adult Amphibian condition has been attained. It is by a secondary adult metamorphosis that the aquatic tadpole turns into the land-traversing frog. The change is not fairly comparable to any process of development by which other animals gain the adult structure characteristic of their class and order, but (in respect of the time of its occurrence) resembles the late assumption of secondary sexual characters, such as the antlers of the stag, or the train of the peacock.

3.—Lastly, we come to the exceptional case of Insects which, unlike other terrestrial animals, develop with metamorphosis. The Anurous Amphibia have prepared us to recognise this too as a case of secondary adult (post-embryonic) metamorphosis. Thysanuran or Orthopterous larvæ cannot differ very widely from the adult form of primitive Insects. From wingless, hexapod Insects, like Cockroach larvæ in all essentials of external form, have been derived, on the one hand, the winged imago, adapted in the more specialised orders to a brief pairing season exclusively spent in migration and propagation; on the other hand, the footless maggot or quiescent pupa.

Insects, like Frogs, disperse as adults, because of the difficulty of the medium, aerial locomotion being even more difficult than locomotion by land, and implying the highest muscular and respiratory efficiency. The flying state is attained by a late metamorphosis, which has not yet become universal in the class, while it is not found in other Tracheates at all. Peripatus, Scorpions, and Myriopods become sexually mature when they reach the stage which corresponds to the ordinary less-modified Insect-nymph, with segmented body, walking legs, and mouth-parts resembling those of the parent.189

The Caterpillar is not, as Harvey190 maintained, a kind of walking egg; it is rather the primitive adult Tracheate modified in accordance with its own special needs. It may be sexually immature, imperfect, destined to attain more elaborate development in a following stage, but it nevertheless marks the stage in which the remote Tracheate ancestor attained complete maturity. Where it differs from the primitive form, hatched with all the characters of the adult, the changes are adaptive and secondary.191

The Genealogy of Insects.

To construct from embryological and other data a chart of the descent of Insects, and of the different orders within the class, is an attempt too hazardous for a student’s text-book.192 A review of the facts of Arthropod development led Balfour193 to conclude that the whole of the Arthropoda cannot be united in a common phylum. The Tracheata are probably “descended from a terrestrial Annelidan type related to Peripatus.... The Crustacea, on the other hand, are clearly descended from a Phyllopod-like ancestor, which can be in no way related to Peripatus.” The resemblances between the Arthropoda appear therefore to be traceable to no nearer common ancestors than some unknown Annelid, probably marine, and furnished with a chitinous cuticle, an œsophageal nervous ring, and perhaps with jointed appendages. Zoological convenience must give place to the results of embryological and historical research, and we shall probably have to reassign the classes hitherto grouped under the easily defined sub-kingdom of Arthropoda.

Sir John Lubbock has explained, in his very interesting treatise on the Origin and Metamorphoses of Insects, the reasons which lead him to conclude “that Insects generally are descended from ancestors resembling the existing genus Campodea [sub-order Collembola], and that these again have arisen from others belonging to a type represented more or less closely by the existing genus Lindia” [a non-ciliated Rotifer].

Present knowledge does not, therefore, justify a more definite statement of the genealogy of Insects than this, that in common with Crustacea they had Annelid ancestors, and that Lindia, Peripatus, and Campodea approximately represent three successive stages of the descent. When we reflect that Cockroaches themselves reach back to the immeasurably distant palæozoic epoch, we get some misty notion of the antiquity and duration of those still remoter ages during which Tracheates, and afterwards Insects, slowly established themselves as new and distinct groups of animals.

CHAPTER XI.