In the above description, which is from my own observations, I have supposed the spider to fix the first and main line of her net to points from one of which she could readily climb to the other, dragging it after her; and many of these nets are placed in situations where this is very practicable. They are frequently, however, stretched in places where it is quite impossible for the spider thus to convey her main line-between the branches of lofty trees having no connection with each other; between two distinct and elevated buildings; and even between plants growing in water. Here then a difficulty occurs. How does the spider contrive to extend her main line, which is often many feet in length, across inaccessible openings of this description?

With the view of deciding this question, to which I could find no very satisfactory answer in books, I made an experiment, for the idea of which I am indebted to a similar one recorded by Mr. Knight1, who informs us that if a spider be placed upon an upright stick having its bottom immersed in water, it will, after trying in vain all other modes of escape, dart out numerous fine threads so light as to float in the air, some one of which, attaching itself to a neighbouring object, furnishes a bridge for its escape. It was clear that if this mode is pursued by the geometric spiders, it would go considerably towards furnishing a solution of the difficulty in question. I accordingly placed the large diadem spider (Epeira Diadema) upon a stick about a foot long, set upright in a vessel containing water. After fastening its thread (as all spiders do before they move) at the top of the stick, it crept down the side until it felt the water with its fore feet, which seem to serve as antennæ : it then immediately swung itself from the stick (which was slightly bent) and climbed up by the thread to the top. This it repeated perhaps a score times, sometimes creeping down a different part of the stick, but more frequently down the very side it had so often traversed in vain. Wearied with this sameness in its operations, I left the room for some hours. On my return I was surprised to find my prisoner escaped, and not a little pleased to discover, on further examination, a thread extended from the top of the stick to a cabinet

or other shelter, and there construct a cell in which the spider remains concealed till the vibrations of a strong line of communication, composed of several united threads, which she has spun from the centre of the net to her cell, inform her of the capture of a fly, to which she then rushes along this bridge. This criticism as to the too extensive generalisation of the procedures of the garden spider above described is perfectly just, as my own observations since the publication of the last edition of this work, but long before I had seen Mr. Blackwall's paper, had shown me. My excuse must be that the observations above recorded (which are left precisely as originally written about the year 1812), having been made on the spur of the occasion in my garden at Drypool, near Hull, when to my surprise I could not find in books any intelligible account of the way in which the geometric spiders construct their nets, were necessarily confined to the common garden species alone found there, and my attention having been subsequently fully occupied in other directions, it did not occur to me that probably the operations of other species might differ from those I had witnessed. These variations, however, do not affect the accuracy of the description above given of the procedures of the species referred to, one of the commonest of the tribe, which description also, except in the two particulars above stated, is generally applicable to the whole geometric race, and has been in great part adopted by Mr. Blackwall in his more full detail of their operations.

1 Treatise on the Apple and Pear, p. 97.

seven or eight inches distant, which thread had doubtless served as its bridge. Eager to witness the process by which the line was constructed, I replaced the spider in its former position. After frequently creeping down and mounting up again as before, at length it let itself drop from the top of the stick, not as before by a single thread, but by two, each distant from the other about the twelfth of an inch, guided as usual by one of its hind feet, and one apparently smaller than the other. When it had suffered itself to descend nearly to the surface of the water, it stopped short, and, by some means which I could not distinctly see, broke off close to the spinners the smallest thread, which, still adhering by the other end to the top of the stick, floated in the air, and was so light as to be carried about by the slightest breath. On approaching a pencil to the loose end of this line, it did not adhere from mere contact. I therefore twisted it once or twice round the pencil, and then drew it tight. The spider, which had previously climbed to the top of the stick, immediately pulled at it with one of its feet, and finding it sufficiently tense, crept along it, strengthening it as it proceeded by another thread, and thus reached the pencil.1

That this therefore is one mode by which the geometric spiders convey the main line of their nets between distant objects, there can be no doubt, but that it is the only one is not so clear. If the position of the main line be thus determined by the accidental influence of the wind, we might expect to see these nets arranged with great irregularity, and crossing each other in every direction; yet it is the fact that, however closely crowded they may be, they constantly appear to be placed not by accident but design, commonly running parallel with each other at right angles with the points of support, and never interfering. Another objection too presents itself. From the experiment related, it is clear that the main line of the net can never be longer than the height of the object from which the spider dropped in forming it. But it is no uncommon thing to see nets in which these lines are a yard or two long fastened to twigs of grass not a foot in height, and yet separated by obstacles effectually precluding the possibility of the spider's having dragged the lines from one to the other. Here, therefore, some other process must have been used.

Both these difficulties would be removed by adopting the explanation of an anonymous author in the Journal de Physique, founded, as he asserts, on actual observation. He says that he saw a small spider, which he had forced to suspend itself by its thread from the point of a feather, shoot out obliquely in opposite directions other smaller threads, which attached themselves in the still air of a room, without any influence of the wind, to the objects towards which they were directed. He, therefore, infers that

1 Some time after making this experiment I stumbled upon a passage in Redi (De Insectis, p. 119.), from which it appears that Blancanus, in his Commentaries upon Aristotle, has related a series of observations which led him to precisely the same result. Lehmann, too, in a paper in the Transactions of the Society of Naturalists at Berlin (translated in the Philosophical Magazine, xi. 323.), has given an explanation somewhat similar of the operations of this very spider, but I am inclined to think erroneous in some particulars. He describes it as emitting numerous floating threads at the commencement of its descent. That he is mistaken in supposing these threads to be more than one, is proved by the fact which I have observed that even that one sometimes breaks by the weight of the spider. How then could an insect almost as big as a gooseberry be supported by a line of the tenuity here attributed to it?

2 An. vii. Vindémiaire. Translated in Phil. Mag. ii. 275.

spiders have the power of shooting out threads and directing them at pleasure towards a determined point, judging of the distance and position of the object by some sense of which we are ignorant. Something like this manœuvre I once myself witnessed in a male of the small garden spider (Epeira? reticulata). It was standing midway on a long perpendicular fixed thread, and an appearance caught my eye of what seemed to be the emission of threads from its projected spinners. I therefore moved my arm in the direction in which they apparently proceeded, and, as I suspected, a floating thread attached itself to my coat, along which the spider crept. As this was connected with the spinners of the spider, it could not have been formed in the same way with the secondary thread of E. Diadema above described.

Probably in this case, as in so many others, we bewilder ourselves by attempting to make nature bend to generalities to which she disdains to submit. Different spiders may lay the foundations of their net in a different manner; some on the plan adopted by E. Diadema; others, as Lister long ago conjectured1, by shooting out threads in the mode of the flying species, as in the instances recorded by the anonymous observer and Mr. Knight. Nor is it improbable that the same species has the power of varying its procedures according to circumstances.

How far these suppositions are correct it is impossible to determine without further experiments, which it is somewhat strange should not before now have been instituted. Pliny thought it nothing to the credit of the philosophers of his day, that while they were disputing about the number of heroes of the name of Hercules, and the site of the sepulchre of Bacchus, they should not have decided whether the queen bee had a sting or not; but it seems much more discreditable to the entomologists of ours, that they should yet be ignorant how the geometric spiders fix their nets. One excuse for them is, that these insects generally begin their operations in the night, so that, though it is very easy to see them spinning their concentric circles, it is seldom that they can be caught laying the foundations of their snares. Yet doubtless the lucky moment might be hit by an attentive observer, and I shall be glad if my attempt to describe their more ordinary operations should induce you to aim at signalising yourself by the discovery. If you failed in solving every difficulty, you would at least be rewarded by witnessing their industry, ingenuity, and patience.

For the latter virtue they have no small occasion. Incapable of actively pursuing their prey, they are dependent upon what chance conducts into their toils, which, especially those spread in neglected buildings, often remain for a long period empty. Even the geometrical spiders, which fix themselves in the midst of a well-peopled district in the open air, have frequently to sustain a protracted abstinence. A continued storm of wind and rain will demolish their nets, and preclude the possibility of reconstructing them for many days or sometimes weeks, during which not even a single gnat regales their sharp-set appetites. And when at length formed anew or repaired, an unlucky bee or wasp, or an overgrown fly, will perversely entangle itself in toils not intended for insects of its bulk, and in disengaging itself once more leave the net in ruin. All these trials move not our philosophic race. They patiently sit in their watching place in the same posture, scarcely even stirring but when the expected prey appears.

1 Hist. Anim. Ang. p. 7.

2 Plin. Hist. Nat. 1. xi. c. 17.

And however repeatedly their nets are injured or destroyed, as long as their store of silk is unexhausted, they repair or reconstruct them without loss of time.

The web of a house-spider will, with occasional repairs, serve for a considerable period; but the nets of the geometric spiders are in favourable weather renewed either wholly, or at least their concentric circles, every twenty-four hours, even when not apparently injured. This difference in the operations of the two tribes depends upon a very remarkable peculiarity in the conformation of their snares. The threads of the housespider's web are all of the same kind of silk; and flies are caught in them from their claws becoming entangled in the fine meshes which form the texture. On the other hand, the net of the garden spider is composed of two distinct kinds of silk; that of the radii not adhesive, that of the circles extremely viscid. The cause of this difference, which, when it is considered that both sorts of silk proceed from the same instrument, is truly wonderful, may be readily perceived. If you examine a newly formed net with a microscope, you will find that the threads composing the outline and the radii are simple, those of the circles closely studded with minute dew-like globules, which, from the elasticity of the thread, are easily separable from each other. That these are in fact globules of viscid gum, is proved by their adhering to the finger and retaining dust thrown upon the net, while the unadhesive radii and exterior threads remain unsoiled. It is these gummed threads alone which retain the insects that fly into the net; and as they lose their viscid properties by the action of the air, it is necessary that they should be frequently renewed.2

1 May not the spinners mentioned by Leeuwenhoek be peculiar to the retiary spiders, and furnish this viscid thread?

2 The accuracy of the fact above stated as to the essential difference between the radii and concentric circles from the presence of globules of gum on the latter only has been denied by the author of Insect Architecture; but as it has been fully confirmed by Mr. Blackwall, and as any one, who will examine a newly-made spider's net with a common pocket lens, and throw a little dust on it, will see for himself what is here described, it is needless to refute an error that has most probably arisen from the examination of old nets, which, after being exposed to wind and rain, often lose the globules of gum from the circles. (Vide Spence in Loudon's Mag. of Nat. Hist. 1832, vol. v. p. 689.)

When the writer of these letters on the food of insects, in examining for himself the whole process, from first to last, of the construction of the nets of the garden geometric spider, observed this remarkable difference between the radii and concentric circles, he had certainly no idea that he had made any discovery, as he never dreamed that so obvious a peculiarity in objects so constantly in view had not been very frequently noticed, and even described, in books, though he had not himself chanced to meet with any such description. But the denial of the fact itself having subsequently drawn his attention to the subject, he is inclined to believe (but without speaking positively on a question which he has not now an opportunity of investigating) that the existence of these gum globules and their peculiar object were first distinctly made known in the present work*; a circumstance which, if the fact

* Dr. Hooke, indeed, in a passage in his Micrographia, p. 202., quoted by Mr. Blackwall (Linn. Trans. xvi. 479.), speaks of the radii of geometric spiders' nets being "all over knotted or pearled with small transparent globules, not unlike small crystal beads or seed-pearls strung on a clew of silk;" but, as he immediately adds, "which, whether they were so spun by the spider or by the adventitious moisture of a fog (which I have observed to cover all those filaments with such crystalline beads), I shall not now dispute;" it is clear that he had no distinct or correct ideas as to the origin of these globules, nor the slightest conception of their use.

In this renewal, as above hinted, the geometrical spiders are constantly regulated by the future probable state of the atmosphere, of which they have such a nice perception, that M. Q. D'Isjonval, to whom we are indebted for the fact, has proposed them as most accurate barometers. He asserts that if the weather be about to be variable, wet and stormy, the main threads which support the net will be certainly short; but if fine settled weather be on the point of commencing, these threads will be as invariably very long. Without going the length, with M. D'Isjonval, of deeming his discoveries important enough to regulate the march of armies, or the sailing of fleets, or of proposing that the first appearance of these barometrical spiders in spring should be announced by the sound of trumpet, I have reason to suppose from my own observations that his statements are in the main accurate, and that a very good idea of the weather may be formed from attending to these insects.

The spiders which form geometrical nets differ from the weavers also with respect to the situation in which they watch for their prey. They do not conceal themselves under their net, but either place themselves

prove to have been so, deserves being held out to the young entomologist in proof how wide a field of discovery must yet remain to be explored, when points at once so curious and yet obvious in the economy of a spider, found in every garden, had so long remained unnoticed.

Another reason for directing attention to this fact is to recommend strongly to comparative anatomists and microscopical observers an investigation of the mode in which the geometric spiders are enabled to spin two different kinds of silk, one gummy and the other not, and whether the spinners noted by Leeuwenhoek, as suggested in a preceding note, are concerned in the process-points to which Mr. Blackwall, in his examination of the spinning apparatus of spiders (Linn. Trans. xviii. 219.), has not adverted. It is obvious that these spiders must either have two distinct sets of spinners, of which one spins the gummy and the other the unadhesive threads, or else, if all the threads proceed from the same spinners, the spider must have the means of passing the threads of the concentric circles through a reservoir of gum so as to stud them with the globules of this substance which give them their fly-catching viscidity. There is, however, a considerable difficulty in the way of this last supposition, for as the threads at their issuing from the spinners are, as has been already explained, so numerous, it is not easy to conceive how, after being united into one, they can be passed through any gum reservoir, nor how, if they were so passed, the gum, instead of being applied to the entire surface of the threads, should come to be divided in the process into distinct and bead-like globules. The subject is certainly highly curious and interesting, and well deserves investigation for an additional reason originally noticed above and confirmed by Mr. Blackwall, that the circular lines differ from the radii and main lines of the net, not only in being studded with gum globules, but in being far more elastic, which elasticity (as well as the viscidity of the gum globules) he found remained unimpaired for more than seven months in a net of Epeira diadema constructed in a glass jar which was placed in a dark closet. (Linn. Trans. xvi. 479.)

Before concluding this long note, an omission in the account of the geometric spiders' forming their nets, in the text, which has been supplied by Mr. Blackwall, should be given, namely, that in the process of spinning the concentric gummy circles, the spider, as she proceeds, destroys the first made distant unadhesive circles which had served her as a scaffolding in placing the former. (Zool. Journ. v. 183.) A curious calculation, also, of Mr. Blackwall's, as to the number of distinct globules of gum in a geometric spider's net, should be noticed. These he found to be 87,360 in a net of average dimensions, and 120,000 in a large net of fourteen or sixteen inches diameter; and yet Epeira apoclysa will, if uninterrupted, complete its snare on an average in forty minutes. (p. 478.)

1 Brez, La Flore des Insectophiles, 129.