Chapter III: Front Matter (3)

Thus, Cannes is from fifty to sixty miles distant from San Remo, but the nests of _N. cæmentaria_, _N. Manderstjernæ_, and _N. Eleanora_ show precisely the same characteristics in either place.

Moreover, the twelve nests referred to were not all taken from one restricted locality at Bordeaux, but were found presenting the same characteristics and occupied by the same spider in three distinct habitats, distant some miles from one another. In two nests several young spiders were found with the mother, and, in one case where the family consisted of twenty-three young ones, I observed that they were not all equally small, and some had nearly attained one-third of their full size.

This agreed with the fact that no very small nests were observed, and it seems probable that the young are not turned out of their nursery quite so early as some of their relations are at Mentone. This, however, varies perhaps in accordance with changes of climate and local conditions.

We failed to detect any other type of nest at Bordeaux than the one described above: and even the cork nests, which we had shortly before seen in such abundance at Montpellier, were apparently absent.

Bordeaux is by far the north-westernmost point in Europe[139] at which any spider constructing a true trap-door nest has as yet been discovered; and the fact that they exist in a climate so different from that of the Riviera and of the whole Mediterranean region, leads me to hope that their range may in reality be much more widely extended than has hitherto been supposed to be the case.

[Footnote 139: Cork nests have however been mentioned as occurring in the neighbourhood of Lyons, which lies in nearly the same parallel of latitude with Bordeaux.]

A glance at the vegetation of this district will suffice to show how little there is that betokens either a warm or dry winter climate; for here the myrtles, oranges and olives are left far behind, and in their place we see tall hedgerow elms, and poplars bearing mistletoe on their branches.

Here therefore we are met by the question, How do these Bordeaux spiders contrive to live under conditions so different from those to which their relations on the Riviera have adapted themselves? How do they bear the cold and damp of the long winter, and how is it that one frail upper door suffices to protect their nest from molestation?

The thick coating of dead leaves, which covered the banks even when we found them, no doubt aids largely in their concealment, and the colder climate probably diminishes the number of their enemies, but their means of subsistence are most likely also less abundant and their period of active life shorter.

The next type we have to consider is a totally new one, and may be distinguished as the _single-door branched wafer nest_. I detected this nest at Montpellier but a few days before the visit to Bordeaux alluded to above.

Circumstances unfortunately prevented me from following up my discovery as closely as I could have wished, and it appears moreover that this nest is far less common at Montpellier than the typical cork nest (_Nemesia cæmentaria_).

I hope therefore that other naturalists will make further investigations, and especially that they will endeavour to secure the male.

I obtained twelve spiders and thoroughly followed the course of ten nests; I opened thirteen more nests, but failed to trace their structure satisfactorily.

The upper part of this nest is shown of the natural size in Plate XVII. with the spider (_Nemesia suffusa_, Camb.[140]) which constructs it. This is again a wafer nest without any lower door, and this absence of a lower door alone distinguishes it as a type from the branched nest represented at F in the diagram, just as the same deficiency separated the Bordeaux type from that at fig. E.

[Footnote 140: We have again in this instance an exemplification of the rule that a new type of nest indicates the presence of a new spider, and hitherto, this rule has proved without exception. Mr. Pickard-Cambridge's description of _N. suffusa_ will be found at p. 295, below. Its slender proportions, cylindrico-ovate abdomen, marked with narrow linear chevrons, and caput without, or almost without, any median line or marking, form some of its more striking characteristics.]

In this new single-door branched type, the branch makes a more or less acute angle with the main tube, and reaches the surface of the ground, but is there closed by a layer of particles of earth slightly bound together with silk, forming an immovable cover or thatch. This cover constitutes, however, but a slight obstruction and could easily be torn away by the spider if she needed to use this passage as a way of escape.

These nests were tolerably plentiful at a place called Les Mourines, a short distance from Montpellier, where they were mixed with cork nests in the steep hedge banks. The nests were from 8 to 10 inches deep, and, as in all the trap-door nests which I have examined, were tenanted by the female alone. It seems strange that this spider, building as she does a nest apparently but poorly furnished either for concealment or defence, should be able to enter into competition with _N. cæmentaria_, whose solid, closely-fitting door appears so perfectly contrived for both. It will probably be found, however, when we are better acquainted with their respective ways of life, that they are really more nearly on a footing than they seem to be at first sight. I detected the remains of ants and the elytra of a beetle in one of these branched single-door nests. Now these may also be found in cork nests, so that _Nemesia suffusa_ evidently competes with _cæmentaria_ for its food, and this is of course the main cause of contention between all living creatures.

It is possible, that, if we knew all the uses to which the branch is put by the spider which constructs it, we should find that the advantages derived in the way of security from the existence of this second passage, counterbalance those possessed by the cork nest, which, though so perfectly closed, has only the one tube, and no other possible way of escape.

It may perhaps be no more than a coincidence, but we can scarcely avoid commenting upon the fact, that, just as this Montpellier wafer nest is simpler in construction than any found along the Riviera, so in like manner is the Bordeaux nest simpler than that of Montpellier. It thus becomes tempting to ask whether, in the case of these wafer nests, we shall not discover that the colder and damper climates are the homes of the builders of the simpler types, while the warmer and drier ones, where more food, more enemies and more competitors are found, are reserved for the architects of the more complicated nests.

Doubtless naturalists will soon discover wafer nests on the slopes of the Pyrenees, as for example at Pau and other winter stations in South-western France; and perhaps the coast of the Bay of Biscay may also yield specimens, even to the north of Bordeaux. If so, this curious speculation as to whether there is any relation between simplicity of structure and warmth of climate, will be put on its trial.

About the very time when I was engaged in digging out these new wafer nests at Montpellier, the celebrated arachnologist, Dr. L. Koch of Nuremberg, had just published[141] an account and figure of a very remarkable nest which he had received from Australia, and which, though differing both in form and proportions from the Montpellier nest, may nevertheless perhaps be referred to the present single-door branched wafer type.

[Footnote 141: Dr. L. Koch, _Arachniden Australiens_, 10te. Lieferuug, Nurnberg, 1874, tab. xxxvii. fig. 3, p. 484.]

This Australian nest, the exact habitat of which is not mentioned, is constructed by a spider now described for the first time under the name of _Idioctis helva_. The nest has a wafer-door about the size of a sixpence, closing a vertical tube less than half an inch long, which meets and opens into a horizontal tube about three inches in length, and forms with it what may be roughly likened to the figure of a capital T inverted, thus, ┹.

The upstroke of the T is however, very short, and one of the arms is longer than the other, and curved downwards at its extremity. This is, as far as I know, the first recorded example of a wafer-nest from the Antipodes, and it may be regarded as one of the first fruits of a harvest which lies ready for the reaping of any naturalist resident in those parts. Hitherto the only nests which I have seen or heard of from Australia were of the cork type (_Ants and Spiders_, p. 132).

Next in order to the single-door branched wafer comes the _double-door unbranched wafer_ type, which is the simplest of all the nests possessing two doors. This habitation, the work of _N. Eleanora_, has been already described (_Ants and Spiders_, p. 106), and I have not much to add to the account there given.

Perhaps some of my readers may remember that, while I was actually engaged on the proofs of _Ants and Spiders_ I had one of these _Eleanora_ spiders in captivity, and that I gave an account (p. 148) of her behaviour up to the latest moment possible. She had been captured on October 23, 1872, and placed, together with five young ones found with her in the nest, on the surface of some earth in a medium-sized flower-pot covered over with gauze. The young ones soon made nests for themselves in the earth, each furnished with its little door, but the mother roamed about on the surface of the soil, and it was not until she had been twenty-one days in captivity that she commenced spinning a silk cell.

This cell in twelve days' time presented the form of a rude figure of 8, and had an aperture at either end; it was just large enough to contain the spider when the legs were extended; its upper surface was attached to the gauze covering of the pot, and its lower to the earth. It was at this stage that the record was broken off, and I will now relate the remainder of the history.

Four days before the cell was commenced, the spider had covered the under surface of the gauze with a semi-transparent film of a substance resembling varnish, which formed a band about three inches long by half an inch wide, close to where the rim of the flower-pot threw the most shade. It was at one extremity of this band that the silk-cell was formed, but it is important to note that this band of varnish was longer than the cell, which only measured an inch and a quarter from end to end, for we shall see that the layer of varnish was apparently laid with a view to further operations.

In four days after the completion of the cell its form was modified, and, during the next ten days (up to December 21st), the spider gradually thickened the walls, and made the form of the cell more and more cylindrical, sometimes closing and at other times opening the extremities.

Between December 14th and 25th, she lengthened out the cell by spinning a cylindrical silk tube in prolongation of one end, and this tube followed the course of the band of varnish, the whole measuring three-and-a-half inches in length by about half an inch in diameter.

It would appear therefore from the correspondence in length between the band of varnish and this silk tube, that she had contemplated the construction of the latter when she first commenced her work on November 3rd.

On January 19th the silk tube parted from the gauze, leaving only the enlarged end which formed the cell still adhering to it. On the following day I observed the very curious fact that when I sprinkled the nest with water, as it was my custom to do every morning, the tube, which had become somewhat flaccid since it had lost its attachment to the gauze, gradually recovered its perfect shape. This was repeated for eleven days, until on the morning of the twelfth day (January 31st), finding the tube completely collapsed, instead of merely sprinkling water over it, I drew a large camel-hair brush loaded with water along its whole length, when the tube started up, and almost instantaneously regained its cylindrical form.

This morning the spider had left her cell, and was roaming about the pot when I wetted the tube, thus proving that she was in no way concerned with its movements, which were no doubt due to hygrometric action.

Between this time and February 25th, I constantly restored the tube to its shape by wetting it in the way above described, but on this day it remained very flaccid, and only expanded partially. For some days previous to this date, the spider had left the tube when it collapsed, and only returned to it again when it had resumed its shape. On the following day I found the entire silk tube and the cell again collapsed and lying flat upon the ground, and this time water failed to produce its previous effect.

The spider then became very restless and excited, and I observed that the door of one of the little nests constructed by one of her five offspring which had been imprisoned in the same pot with her, had been torn off, and thrown on one side, and there could be little doubt but that the mother had been guilty of this very un-maternal action. By the evening she had pulled up her collapsed tube from its attachment to the earth, and had coiled it in a confused heap. Seeing this, and fearing that, in her distress and excitement, she might do further damage to the young spiders, which had up to that time thriven well, I made a cylindrical hole for her in the earth, supposing that she would at once take possession of it. On the following morning, however, the mother spider had advanced some way in building another figure-of-8 cell, rising the shrivelled silk of her previous dwelling as a foundation.

In twenty-four hours this second cell was complete, and closely resembled the former one, save that the smaller end of the 8 was turned in the opposite direction, but, on examining it, I found to my surprise that it was empty! The spider had taken possession of the hole I had made for her, which she had at first refused to notice, and was busily employed in lining it with silk and furnishing it with a covering composed of silk with earth and fragments of moss woven into the surface. By mid-day the aperture was completely closed, but there was no moveable door. From this time (February 28) up to April 12, the spider lived in this hole, which she eventually furnished with a distinct wafer-door, and, as I found on opening the nest, with a typical lower door also. This latter was not neatly made, but still it possessed all features the essential which characterize these lower doors in the nests of _N. Eleanora_.

So this captive _Nemesia Eleanora_ lived in a flower-pot in my bedroom for more than five months and a half, during which time she absolutely refused to burrow or to attempt any kind of excavation, but passed the greater part of that period on the surface of the earth in a silk tube ending in an oblong enlargement, utterly unlike her normal habitation. Finally, when I had done the digging for her, she furnished the cylindrical hole I had bored in the earth with a silk lining, and made it secure with her own two typical doors.

The figure-of-8 cell which she constructed at first, and subsequently modified until it became the oblong enlargement of the tube alluded to above, was totally unlike any form of trap-door spider's nest known to me; but in its ultimate shape (which resembled that of the glass part of a thermometer with an oblong bulb, save that it was curved and not straight), I think we may trace some resemblance to the silk tube which is made by _Atypus_, and of which a figure is given at A, Plate XIII., p. 183; the mouth of the tube made by my captive was, however, open. It is curious, also, when we recall this resemblance, to note that Mr. Brown has recorded, in his observations alluded to above (p. 185), that the tube of one of the nests of _Atypus_, which he brought home in a collapsed state, showed a somewhat similar tendency to become distended. For, on opening the box in which they had been carried, he perceived a movement throughout the tube as if it were becoming inflated, and though this inflation appeared to subside shortly after, yet the following morning the tube had recovered its cylindrical shape. I am tempted to believe, though this is mere conjecture, that the box in which these tubes were put contained moisture, and that their apparent inflation was due to the same hygrometric action which, was displayed in the tube of _N. Eleanora_. I regretted that I was unable to continue my observations on this captive spider, as it would have been interesting to know how long she would have lived contentedly and in good health under the conditions described above, but I left Mentone at the end of April, and was unable to take her alive with me to England. When removed from her nest in the pot on April 12, she appeared in perfect condition, and I placed her in a hole which I made for her among some stones in a garden at the back of the house, hoping to find her again on my return to Mentone in the autumn; this hope was, however, not destined to be realized.

I shall, however, have occasion to speak again of the young captives of this species (_N. Eleanora_), in the concluding remarks which will follow these detailed accounts of the nests and their occupants, when the behaviour of captive trap-door spiders generally will be treated of.

The next type of trap-door nest is one to which I have found it difficult to assign a descriptive name, and I am compelled for the present to speak of it as the _Hyères double-door branched wafer_ nest.

One of its most distinctive features is found in the shape of the lower door, fig. F 1, Plate XIV., and figs. A 1, A 2, Plate XVIII., which may be said to be double, presenting two crowns, one of which fits into the main tube and the other into the branch, but I could not see my way to employing this character in naming the type. The nest is, however, quite distinct from all the others, and is inhabited by a new species of trap-door spider (_N. congener_, Camb.[142]). The characteristic portions of this nest are shown in Plate XVIII., and fig. A 3, in the same Plate, represents its occupant.

[Footnote 142: Mr. Pickard-Cambridge's description will be found at p. 292, below. In its characters this female spider (the male is unknown) most nearly resembles _N. cæmentaria_, but differs, among other points, in markings and in having one or more spines on the genual joint of leg, these spines being almost always absent in the same joint in _cæmentaria_. The nests of the two species are totally unlike.]

The hedge-banks near Hyères, and also about the railway station of the same name, which is some 4 miles from the town itself, are frequently tenanted by this spider. During a short stay there in May, 1873, I secured a large number of specimens, and verified the structure of the nest by a careful examination of thirty-eight examples. The nest is invariably branched and furnished with a lower door, but the branch is of variable length, usually short, and never, as far as I could detect, quite reaches the surface. In some cases this branch was so short that it could scarcely contain the spider, and, under these circumstances, it is not easy to conceive any other use for it than that of retaining the lower door when not in use. It may, however, enable the spider to take up a rather better position when engaged, as she frequently is if disturbed, in keeping the main tube closed by pressing the lower door upwards with her feet, for then her head points downwards, and her abdomen rests in the branch.

I have seen her in this attitude on several occasions when I had cut out a block of earth similar to that figured in the plate. The lower door is quite unlike that of either of the other two double-door wafer nests, being wedge-shaped, tapering from below upwards to the hinge, which is always placed at the point of bifurcation of the tubes, and having two crowns separated from each other by the gusset-like web of silk which connects the door on either side with the lining of the main tube, one of these crowns fitting into and closing the main tube, while the other fits into the aperture of the branch.

The wedge-shaped structure of the door is seen in its most exaggerated form in the nests of the younger spiders (figs. B, B 1, Plate XVIII.), and becomes less so in the older and larger ones (figs. A 1, A 2). I have even seen some of these lower doors, evidently made by old spiders, which were so much flattened as to bear a considerable resemblance to that of _N. Eleanora_.

The main tube of the nest is from 10 to 12 inches long, and usually enters the earth almost horizontally, bending downwards from the point at which the branch joins it, and where the lower door is hung. This causes the lower door to lie nearly horizontally when not in use, and its lower crown probably serves, by fitting into the aperture of the branch, to sustain it in this position and prevent it from falling forward. The point of bifurcation is placed, as a rule, much nearer to the entrance of the nest, than it is in the two other branched nests, and occurs usually within two inches of the surface of the earth; so close is it indeed that, on lifting the upper door and looking in, one may frequently see the lower door move across and close the passage down the main tube, pushed by the spider from below. This frequently enabled me to secure the spider without having to follow her to the bottom of the nest; and, when fortune favoured me, I secured a block of earth by one rapid sweep of the knife (a common table-knife), which furnished me at once with a good specimen of the nest and of its occupant.

When the spider has once fairly determined upon resistance, it is scarcely possible to make her retreat without destroying the nest, and, in one case, when I tried to push the lower door down from above, while she was pressing it upwards from below, I found that, without crushing my opponent, I could not succeed.

There were probably young in the nest on this occasion, for I have frequently found them in the nests with the mother at this season. In no case did I even catch a glimpse of the male, and this sex is at present unknown.

The young spiders make their nests at an early age, and there can be no doubt that _N. congener_ enlarges its dwelling from time to time as growth demands, just as the trap-door spiders at Mentone do. Indeed in one of these new Hyères nests I found, outside the main tube and some way above the existing lower door, a former and disused lower door much smaller than the one then in use, and which had evidently belonged to the nest at a previous stage of its development. I have observed this before in the nests both of _N. Manderstjernæ_ and _N. Eleanora_.

This new type is strictly intermediate between the double-door unbranched wafer nest constructed by _N. Eleanora_, and the double-door branched wafer with the descending cavity which I am now about to describe.

This latter nest, the work of _N. Manderstjernæ_, Auss.[143] (formerly called _N. meridionalis_), has already been partially made known by the figures and description given of it in _Ants and Spiders_ (Plates IX., X., and XI., pp. 98, 100, and 104); but I have to confess, with great regret, that when these illustrations and descriptions were published, I was not fully acquainted with the true structure of this nest, having overlooked the existence of a short descending cavity which leaves the main tube a little above and on the opposite side to the ascending branch. This cavity is always present, but the very largest and oldest spiders usually allow it to become filled up with remains of food and particles of earth, and sometimes even spin silk across its entrance, in which case it can only be traced on very close examination.

[Footnote 143: This spider was described by Mr. Pickard-Cambridge at p. 101 in _Ants and Spiders_, under the name of _N. meridionalis_, Costa. This name has now to be abandoned for reasons given in full by Mr. Cambridge at p. 283, below. It would appear that a spider discovered by M. Simon in Corsica corresponds more closely with the _N. meridionalis_ of Costa than our spider of the Riviera does. Moreover, since _Ants and Spiders_ was written I have had the good fortune to obtain at Mentone four male examples of our supposed _meridionalis_, and these prove to possess the same characters as those assigned by Prof. Ausserer to a male spider which was captured at Nice, and named by him _N. Manderstjernæ_. This specimen is now in the possession of Dr. L. Koch, to whom I am much indebted for having kindly entrusted it to me for examination. This enabled Mr. Pickard-Cambridge to assure himself of the specific identity of his _N. meridionalis_ with _N. Manderstjernæ_, which latter name it must for the future bear.]

It was from an old nest such as this, in which the descending cavity had been closed up, that the large drawing at fig. A on Plate IX. of _Ants and Spiders_ was made, and this figure, therefore, still remains substantially correct.

But in the case of the other illustrations--namely, fig. B, Plate IX., fig. A, Plate X., and figs. B and B 1, Plate XI., where nests of young spiders, or of spiders which, though adult, have not attained the maximum size, are represented, this descending cavity, though overlooked by me, should have been shown, for it must certainly have existed.

Its presence was first observed by the Honourable L. G. Dillon, who detected it when tracing the course of the main tube upwards from below. I had always followed the tube from above downwards, and in so doing must have unwittingly filled up the descending cavity (the existence of which I was far from suspecting) with detached particles of earth.

I will own that, when Mr. Dillon first showed me this new feature, I hoped that it might prove to be something accidental and exceptional; and it was only after careful examination of a large series of nests of all sizes, that I gradually and almost unwillingly admitted that this descending cavity formed an important feature in the typical structure of the nests.

I now see, however, that the presence of this cavity adds considerably to the interest of the structure as a whole, and places its architect quite at the head of all the builders of trap-door nests. This type should now be called, for the sake of distinction, the _double-door, branched, cavity, wafer_ nest, to avoid confusion with the _Hyères branched nest_.

I am now about to endeavour to atone for my past oversight by giving new illustrations (Plate XIX., figs. A and B) and descriptions of this very remarkable nest; while I would at the same time beg the indulgence of my readers for past and present shortcomings, reminding them that the interest which attaches to structures of this kind is proportioned to the complexity and subtlety of their contrivance, and, therefore, to the difficulty we experience in properly understanding and describing them.

It will be seen by a reference to Plate XIX.,[144] figs. A and A 1, that in addition to the cylindrical branch, which mounts upwards, there is a shorter branch which leaves the main tube on the opposite side (on the left as seen in the Plate), and takes a downward course. Now this descending branch, which is barely more than an inch in length, is a cavity of variable form, being sometimes cylindrical, and sometimes egg- or even watch-shaped,[145] but there is one particular in which it never varies, and that is the position of its elliptic orifice. This orifice is always situated on the opposite side of the main tube to that on which the ascending branch leaves this latter, so that the whole nest, when seen in section, presents the figure of a St. Andrew's cross, only with arms of unequal length.

[Footnote 144: A nest of a scarcely half-grown spider is here represented in order that sufficient space might be gained to show the lower door in its two positions. The perfect cavity is still found in nests of much larger dimensions, and occasionally, indeed, in nests of almost the maximum size.]

[Footnote 145: I take the liberty of coining a word to replace "lenticular," the form of a watch being more familiar than that of a lens.]

But the most remarkable point is that, when the lower door is pushed across so as to close the main tube (as shown in fig. A, Plate XIX.), it will invariably be found to lie in such a position that its lower extremity exactly meets the lower lip of the orifice of the descending cavity, when it will be seen that the semi-cylindrical surface of the lower door then coincides with, and appears to continue and form part of, the lower wall of the descending cavity on the one side, and of the corresponding wall of the main tube on the other. When the upper portion of the main tube is thus united to the cavity the two combine to form what appears like a short, independent unbranched nest.

Now, if we fancy ourselves an insect entering the nest in search either of the spider, her eggs, or young, I think it is plain that, when the lower door is in this position (fig. A), we should probably walk straight down to the bottom of the cavity, expecting to find our prey there, and should then return by the way we came, impressed with the belief that we had explored the whole nest, the secret of the lower door remaining undiscovered.

Whether this imaginary case may, or may not, represent what really takes place, is of course mere conjecture; but the constant occurrence of this beautiful adaptation of the various parts to one another, surely points to the conclusion that this is no mere coincidence, but rather a subtle contrivance having some very definite use and meaning.

We must admit, however, that it is difficult to conceive why, if this structure is of such great utility, it should be abandoned by the oldest and largest spiders.

Among the possible answers to this question I think that one of the more probable is that this arrangement may have been specially devised for protection against some enemy which the aged spiders have ceased to fear.

Indeed it is not unlikely that these aged spiders may have come to a time of life when they no longer lay eggs, and so do not need to keep up all the defences which they employed when they had families to protect.

Since my attention was drawn to the existence of this cavity in the dwellings of _N. Manderstjernæ_ I have never noted the presence of young in those nests in which the cavity was filled up and disused; but then I have only exact records with reference to this point in the case of seven nests.

In these seven nests, however, there was no free cavity, and there were no young spiders, though it was at the season when it was common to find young in the nests.

The question, therefore, remains open, and further observations on this head would be very acceptable. I detected the _débris_ of insects, and especially the horny coats of ants, in the descending cavity, in many nests; and in some of the oldest, where it had become completely blocked up, these remains still indicated its former outlines and position.

The nests of _N. Manderstjernæ_ at Cannes correspond both in respect of the cavity and of their other characteristics with those at Mentone. _N. Manderstjernæ_ occurs pretty abundantly at San Remo in the olive-grounds east of the Sanctuary, but I can say nothing as to whether the nests there possessed the cavity or not, for, when I was there, I was not aware of its existence. I obtained a single example of _N. Manderstjernæ_ and its nest at Hyères, and this is the westernmost point at which this species has as yet been detected.

We have now passed in review all the seven known types of true trap-door nest, and have taken note also of the lower and more rudimentary forms of nest, such as that of _Atypus_, and the funnel nest of _Cyrtauchenius elongatus_, neither of which is furnished with a door.

Among the true trap-door nests, those of the cork type stand in a measure alone, being distinguished from all the others by their solid surface doors, composed of many layers of silk and earth; and we do not at present know of any intermediate forms linking the cork and wafer types together. But among the various nests which represent the wafer type the case is different, for here the types naturally fall into a progressive series, such as that represented in the diagrams (Pl. XIV., p. 193).

If we try to picture to ourselves the stages through which the most complicated wafer nest--namely, that of the _double-door, branched, cavity_ type (Diagram G 1) may have passed in the course of its development from a simpler ancestral form, we should _à priori_ expect to find precisely such structures as the _Hyères double-door branched_ nest (Diagram F), and the _single-door branched_ nest (Diagram D) forming successive halting-places in the advance from the primitive _single-door, unbranched_ nest (Diagram C).

The _double-door unbranched_ type may in like manner find its prototype in the same original single-door unbranched nest (C), which we may look upon as the parent idea, from which all these structures have been derived.

Bearing this in mind, and remembering that kinship between living creatures is not only revealed to us by likeness in structure and colour, but also by similarity in habits and instincts, it becomes of interest to trace any resemblance that may exist between these wafer-nests and the dwellings constructed by _Lycosa narbonensis_, a species belonging to the allied family of _Lycosidæ_, and which closely resembles the true _tarantula_[146] of Southern Italy.

[Footnote 146: In the United States, and indeed in the New World generally, it seems to be the custom to call all the larger "ground spiders," and especially the trap-door spiders, Tarantulas, but these, in fact, form a distinct group by themselves, belonging to the family _Lycosidæ_.]

I first made the acquaintance of _Lycosa narbonensis_ near the glass-works west of Cannes, where this spider may not rarely be found living in tubular burrows in sandy clearings among the pine woods along the shore (_Pinus pinea_, the stone pine).

I have already (_Ants and Spiders_, p. 146), alluded to an account given by M. Léon Dufour of his observations on the nest and habits of the true tarantula (_Lycosa tarentula_), which he discovered in Spain.

The nests of _L. narbonensis_ at Cannes resembled those described by M. Dufour, but the cylindrical, subterranean burrows were apparently shorter. It was extremely difficult to trace their course, on account of the loose sand which poured into the tubes and choked them up, and I only succeeded in doing so completely in one case, when I stuffed the tube with cotton-wool before proceeding to dig. Here the open tube, which was quite simple, and about 1 inch in diameter, descended vertically for 3-1/4 inches, and was then suddenly bent so as to become horizontal, terminating shortly afterwards in a triangular chamber, the floor of which measured 2 inches across at the widest part, and was strewed with the remains of beetles and other insects.

The nest was lined throughout with coarse silk, which had a blackish hue, owing to the presence of the filaments of what I believe to have been some undeveloped fungoid growth. The mouth of the tube was open, and frequently surmounted by a short tubular prolongation, commencing at the surface of the ground, which formed a sort of chimney about an inch high and from an inch to an inch and a quarter across; this was composed of fibres of plants, pine-needles, and especially of a large branching lichen, very common in the neighbourhood of the nests, and all these materials were woven together and kept in place by a few threads of silk spun here and there.

It was not every nest that was furnished with a chimney, nor were all the chimneys equally complete, for in some cases they consisted merely of a small rim or one-sided lip, while in others they resembled little birds' nests, and were sufficiently firm and compact to permit of my carrying them away. It appeared to me that these chimneys served as screens to prevent the loose sand from being swept into the burrows by the winds which rage over that open seashore plain, and that they were more or less complete in proportion as the exposure was greater or less, and the sand looser or more bound together.

I captured eight of these spiders, and here, as in the trap-door group, the female alone inhabited the nest.

Besides this habit, they have other points in common with trap-door spiders; such, for example, as the resemblance which exists between this nest and that of _Theraphosa Blondii_ from Brazil (see p. 188, above), and between the chimney of this Tarantula and the aërial prolongation of the tube sometimes found in nests of the wafer type.

But perhaps the most suggestive point of resemblance consists in the habit which this Tarantula possesses of covering and closing the aperture of the nest during the winter with a thin layer of materials, similar to those of which the chimney is composed, and, like them, bound together with silk. This is, in fact, an immovable wafer-door, and precisely resembles those which I have seen constructed by _Nemesia Manderstjernæ_, and _N. Eleanora_, when captive and placed in an artificial hole in the earth.

The tubes are, as has been already stated, open during the spring, and we may suppose that the spider, on the approach of warm weather, wakes up from her winter lethargy, and tears away this concealing thatch. But if one of these spiders should by chance happen to free this silk-woven thatch by cutting round some three-fourths of its circumference, so as to leave it still attached to the rim of the aperture of the nest by the remaining quarter, she would then have made for herself a veritable, though rather rude trap-door of the wafer kind.

It is most likely, however, that the spider knows what she is about and that a door to her dwelling would be the reverse of an advantage to her, for she is more powerful and swifter than the generality of European trap-door spiders, and, as she probably lives by leaping out upon and hunting her prey, she no doubt needs to have the entrance to her nest free of all encumbrance.

I am indebted to the Rev. W. G. Brackenridge for evidence of the very interesting fact that _Lycosa narbonensis_ closes her nest at Cannes in the winter.

I was aware that Latreille stated that the Tarantula possessed this habit,[147] and I was anxious to know whether the species which I had detected at Cannes, inhabiting as it did open nests in the month of May, would also exhibit this curious custom. Being unable to visit Cannes myself during the winter, I applied to Mr. Brackenridge, who, on the 28th of January last (1874), secured a very perfect specimen of the aërial portion or chimney of one of the nests having the orifice closed in the way above described, and most kindly transmitted it to me.

[Footnote 147: P. A. Latreille, Mem. Soc. Hist. Nat., Paris (an. VII. de la République), p. 124: "L'araignée _tarentule_ ferme aussi son habitation, mais cet opercule n'est pas mobile, et n'est construit que pour l'hiver."]

I have, on a very few occasions, found the doors of a wafer or cork nest spun up during the winter at Mentone, and on digging have discovered the spider alive, though partially torpid, inside; but this is quite an exceptional event. I should much like to know, however, whether this becomes the rule in the case of the nests of those trap-door spiders which inhabit climates less favoured than that of Mentone.

In my concluding remarks in _Ants and Spiders_ I called attention to the importance which attaches to a knowledge of the food and manner of feeding of any creature whose life-history we may wish to study, and I would now once more press the subject on the attention of my readers. For the range and distribution of a species largely depends upon the nature of its food, and this will also be an indication of the rivals with which it has to compete in the struggle for existence; the times and seasons of its activity, and in many cases even the structure and position of its dwelling-place will be governed by this same all-important question of food-supply.

I have now detected the remains of insects, and of ants especially, in the nest of every species of trap-door spider which I have examined _in situ_; very frequently, however, one may open several nests in succession without finding any of these _débris_, and at other times they will only be detected beneath the existing bottom of the tube, layers of silk having been spun over successive layers of refuse.

The horny coats of ants form by very far the largest proportion of these remains, and I have lately been much struck by the number of instances in which, while digging out ants' nests at Mentone, I have found trap-door nests (especially those of _N. Manderstjernæ_ and _N. Moggridgii_) in their midst, the tubes often traversing the very heart of the ants' colony and coming into close contact with the galleries and chambers of the ants. The doors in these instances had almost always escaped my notice, and, indeed, they so closely resembled the surface of the ground that even when I knew, from having accidentally cut across the tube below ground, that one of these doors must lie near a given spot, yet I could only discover it by following the passage from below upwards. This perfect concealment is doubtless of essential importance to the spiders' success in life, for, if they once alarmed the whole colony of ants and let them know the exact whereabouts of their lurking-place, they would soon learn to avoid it.

But, as it is, the work of opening the door, snatching in an ant, and closing it again, is but the affair of a second or two, and before the companions of the victim have time to realize the nature of the phenomenon, the gaping earth has closed again and become once more, to all appearance, part of the solid and trustworthy ground.

I have seen _N. Manderstjernæ_ snatch at insects in this way during the daytime, and I well remember how I started on one occasion when, as I was looking fixedly at a small blue gnat which I had taken for a moth, I saw the earth suddenly open and one of these spiders partly emerge, make a swift stroke at the insect, and withdraw again as swiftly.

I have found the remains of ants, of beetles of many species and different sizes, of wood-lice (_Oniscus_), and of earwigs (_Forficula_) in the nests of _N. Eleanora_ and _N. Manderstjernæ_, and the wings of a large green field-bug in the nest of the former. I have only once detected traces of food in the dwellings of _Cteniza Moggridgii_, and these consisted of minute fragments of the integuments of insects, none of which were certainly recognisable, though I believe that they partly consisted of the coats of a small species of ant. The rarity or complete absence of the wings of insects which habitually fly rather than crawl on the ground, and my inability to discover either snares or any evidence that these spiders ever leave the nest, lead me to believe that they live (at any rate from October to May) by dragging into their nests any insects which approach within reach.

Ants, earwigs, beetles, and wood-lice are precisely the very creatures which would fall a prey to the spider without obliging her to leave her nest, and it is accordingly their remains that we find.

On one occasion, however, at Montpellier, my sister detected _N. cæmentaria_ in the act of devouring a fair-sized caterpillar, to obtain which there is some reason to think she must have left her nest. We were out together on the 8th of May last (1874), hunting for the new wafer nests of that district, under the kind guidance of M. Lichtenstein, when my sister called our attention to a caterpillar, the body of which partly projected from the tube of a cork nest (_N. cæmentaria_), and prevented the lid from closing.

On closer examination we found that the spider was in the act of devouring the caterpillar, and had already sucked out the juices from the anterior portion, while the middle and posterior parts of the body still resisted, and the legs clung tenaciously to the lip of the nest.

M. Lichtenstein told us that this larva, which when entire must have been rather more than an inch long, was that of the mullein moth (_Cucullia verbasci_).

It was not full grown, and as there were no mullein plants within some two feet of the nest and this caterpillar will not leave the plant on which it feeds unless compelled, it would seem as if the spider must have gone afield in order to capture it. It is possible, nevertheless, that the caterpillar may have fallen within reach of the spider when blown off the mullein leaves by the wind.

I have, unfortunately, but few details to give of the nocturnal habits of the trap-door spiders. It would appear, however, that they are more active by night than by day, and that it is more common to find their doors ajar at night, with the spiders posted on the look-out at the narrow opening. This is borne out by my observations on captive spiders, to which I shall allude shortly.

When at Hyères on the 11th of May, 1873, the evening being very warm and a bright moon shining, I went at 8:30 P.M. with my father and sister to see what the spiders would be doing on a hedge bank where we had previously marked five cork and eight wafer nests. The moonlight did not fall upon this spot, but I was provided with a lantern, and by its light the nests at first appeared to be tightly closed, but we soon perceived first one and then another with the door slightly raised, ready to close on the smallest alarm, whether from a footfall or from the flickering of the lamp. When the light of the lantern was steady it did not appear to frighten the spiders in the least, even when brought to within a few inches of the door,[148] and this enabled me to watch them very closely. On either side of the raised door of one of the wafer nests I could see the feet of the spider projecting, and just at that moment I caught sight of a beetle close at hand, feeding on the topmost spray of some small plant below. Using every precaution, I contrived to gather the spray without shaking off the beetle, and gradually pushed it nearer and nearer to the nest. When it almost touched the lip of the nest the door flew open, and the spider snatched at the beetle and dragged it down below.

[Footnote 148: This had been observed before both by my father and Mr. Dillon when watching the trap-door spiders at night at Mentone.]

For a few seconds the door remained tightly closed, and then, to our great surprise, was suddenly opened again, and the beetle was cast alive and unharmed out of the nest. I immediately secured the insect, which proved to be the common _Chrysomela Banksii_.[149]

[Footnote 149: I am indebted to Mr. F. Smith for the name.]

I cannot doubt that this beetle was distasteful in some way to the spider, for it was neither so large nor so powerful as many beetles the remains of which I have found in the spiders' nests, and, besides, it did not escape from the nest, but was distinctly rejected by its captor.

This shows that this spider does not know instinctively what insects to reject and what to take.

This little episode was scarcely ended when I espied a wood-louse (_Oniscus_) walking down the bank, not far from another of these wafer nests. By a little guidance I managed so to turn its course that this unsuspicious crustacean went straight to the very point I wished, and made as if it would walk over the spider's door; but no sooner was it well within reach than, quick as thought, the spider clutched it and dragged it in. No rejection followed on this capture, and, though I could not actually witness the conclusion of this adventure, I do not doubt that it ended in a tragedy and a supper.

In these two cases, as in all those previously noted, the spiders did not leave the nest nor allow the door to close behind them, but kept it propped up on the abdomen and hindmost pair of legs. In this way the act of seizing their prey, and that of withdrawing into the nest, were almost simultaneous.

In no case did we see any of these spiders out of their nests, and their behaviour by night appeared to be the same as by day, only that they were bolder and more on the alert.

The spiders in the cork nests (_N. Moggridgii_) resisted our attempts to raise their doors just as rigorously as in the daytime.

All the spiders which I have kept in captivity have shown themselves more active at night than during the day, and I imagine that experience has taught them that fewer of their enemies are then abroad, while ants, beetles, wood-lice, and other creatures upon which they prey are quite as nocturnal as themselves.

I brought back to England some young cork and wafer spiders from Hyères, and one adult cork (_N. Moggridgii_). The latter was placed in a small tin box, with moss and a little earth at the bottom, on the evening of May the 10th, 1873, and by next morning she had made a silk tube through the moss, carrying up earth from below for the purpose of strengthening its walls on the outside. On the 13th of May the tube was furnished with a perfect door.

I hoped that this spider might lay eggs in her prison,[150] and therefore broke up her nest from time to time after my return to London in order to search for them. Between the 27th of May (when her nest had been transferred into a box of earth) and the 6th of October I destroyed her dwelling four times, and after each demolition she furnished the cylindrical hole which I bored for her with a lid, having thus made five doors since her capture. I got no eggs however, though the spider appeared in perfect health.

[Footnote 150: Strange to say, though I have opened so many nests at different seasons of the year, and found young apparently quite recently hatched, I have never been able to find the eggs of a trap-door spider.]

Neither this spider nor the true _N. cæmentaria_ of Montpellier appears to have any idea of digging a hole when placed on soft earth if they are adult; and the same thing is true of _N. Manderstjernæ_ and _N. Eleanora_, but the young of all these spiders readily excavate nests for themselves.

I have once seen a nearly full grown, and probably adult, _Cteniza Moggridgii_ make a perfect tube and furnish it with a moveable door in a single night when confined under gauze on moist earth, but this is the only instance (except that of _Cteniza Californica_, recorded above) in which I have known an adult trap-door spider excavate or attempt to do so.