to think that L. usitatissimum is but an annual form of L. perenne, so that this year I shall collect the seeds of my perennial patch with a view of commencing an annual cultivation. At all events, should I fail in proving this point, we may fairly expect other changes of great interest, seeing that so much has already been done in bringing a little straggling linseed from its wild habitat, and cultivating in a different soil and climate, not by imitating its wild conditions, but by making for it a new soil, and planting in rows so that one row has the effect of inducing the upright growth to its neighbour, -a fact readily seen in examining the growth of my plant as its shoots first start in a trailing method-a circumstance which shows that in order to test the capabilities of some plants for a crop, we can only do so not by growing single specimen examples, but by planting a quantity side by side.

As subjects for experiment, it fortunately happens that the linseeds are readily affected by cultivative processes, so that we possess in them subjects capable of affording much information as the result of carefully conducted experiments, which leads me to remark that, as there are some tribes of plants which we cannot so easily act upon, permanency of our appointed species must not be concluded from the failure of our limited experiments, though, on the other hand, species must give way in those cases where as the result of properly conducted experiment the seed of one plant can be made to produce what has been considered as a distinctly specific form.

As regards medicinal plants, such specimens as Hyoscyamus, Datura, Papaver album, Coriander, and Caraway seem to do remarkably well in a not over-good soil and with but little trouble, so that where a market can be got for the produce, it might be worth while to attend to their cultivation, especially in corners.

I shall here only remark upon experiments with the Datura Stramonium and D. Tatula. A plot of each of these species was sown side by side, the former from seed grown in the district; the latter from seed kindly communicated by Mr. Savory the eminent apothecary and chemist, of New Bondstreet. Of the former not one seed came up, whilst of the latter several plants at the time of my writing are in great perfection. I am informed by Mr. Savory that this species is highly valuable as a remedial agent, it being much more active and uniform in its action than the D. Stramonium; and he recommends it in the shape of cigars. Though these plants have been referred to under distinct names, there can, I think, be but little doubt that they are only varieties. The flowers of my specimens are but very slightly tinctured with purple. These plants are very abundant in the United States, the tinctured variety being much more common towards the South than in the Northern States, and it is not at all improbable that the want of colour in my specimens is the result of the cold, exposed climate of my garden, and poor soil in which I have planted them*.

7. WEEDS.-In this class I would notice the following plots:—a. Allium vineale; b. Carduus acaulis and others.

a. A plot was planted in the spring of 1856 with young plants of Allium vineale with the view of showing my class its method of growth, I pointing out to them how to get rid of so direful a pest. In the summer it had grown to good flowering heads, when, fearing lest it should overrun the garden, I had them pulled up and put into a weed fire to burn. The plot was left untouched until the spring of 1857, when to my astonishment young plants shot up, and the rows of this plot were as complete as in the former season. Upon reflection I saw in this a lesson which I had not my*Beck in his United States Botany' gives the D. Tatula as a variety of D. Stramonium. The former is called the Indian, and the latter the American thorn-apple.

self sufficiently studied; in order to explain which it will be necessary to point out that around the bulb of this plant, will be found from one to four bulblets, which at the time the plants begin to dry are easily separable from the parent: it therefore happened that upon pulling up the stem, the bulblets became detached and caused a thicker plant to spring up where I had thought it destroyed. This shows how even the pulling of a plant of this character is inefficacious for its destruction; and it may further be appealed to as one of those accidental experiments which almost every plot presents, for it may be observed that in these plots many facts (of agricultural interest especially) are daily unfolded by the College Garden experiments, that I have not commented upon in this report.

As regards the Carduus acaulis, it will here only be necessary to say that having found a new locality in Wilts. for Carduus tuberosus, I have brought a few specimens into my garden, and as will be seen from a separate paper which I have laid before the Section on this discovery, I have an idea that the C. tuberosus is but a hybrid. I am cultivating the C. acaulis and C. acanthoides side by side, in the hope of being able to prove this by experi

ment.

8. FLOWERING AND ORNAMENTAL PLANTS.-These for the most part consist of such specimens as may be of use for teaching, or ornament in the lecture-room; and many of them afford interesting examples of departure from recognized typical forms as to be of value in teaching, whilst others seem to grow wildly and lose their whole cultivative characters. As yet I have not attended to the cultivation of flowers merely as illustrations of transmutation of species; but I am convinced that such genera as Primula, Viola, Myosotis, and Malva, &c., would furnish a vast amount of interesting matter as the result of time and attention bestowed on their investigation.

Here then, for this meeting, must end my notes; if, however, the Section should deem them, or the class of experiment they have reference to, worthy of continuation, the subject offers a field sufficiently wide, and, I think, important for much future investigation and description, as it appears to me that it is upon the noting and collecting such facts as can only be obtained where the subjects of them are under constant observation, that we can hope for much light being thrown upon the at present obscure subject of specific distinctions; and here, whilst experiments are being made upon this matter, it is not too much to state that other facts of great interest are constantly presenting themselves, so that while we are collecting evidence of a scientific kind we may also expect to make experiments tending to useful practical and economic discovery.

On the Resistance of Tubes to Collapse.
By WILLIAM FAIRBAIRN, F.R.S.

Ar the joint request of the British Association and the Royal Society, a series of experiments was undertaken to determine the laws which govern the resisting powers of cylindrical tubes exposed to a uniform external pressure, and from them to determine their strength, and deduce rules for proportioning the internal flues of boilers and similar vessels.

Hitherto it has been considered as an axiom of boiler-engineering, that a cylindrical tube placed in the position of a boiler flue, was equally strong in every part when subjected to a uniform external pressure, the length not

affecting the strength of a flue placed in such circumstances. This rule is, however, applicable only to tubes of infinitely great length, or to tubes unsupported by rigid rings at the extremities; it is very far from true where the length of the tube does not exceed certain limits, and where the ends are retained in a cylindrical form by being securely fastened in rigid frames to prevent their yielding to external pressure. Some experiments upon large boilers, with flues 20 to 30 feet long and about 3 feet diameter, first led to misgivings on this subject, by indicating the greater strength of the shorter flue. This anomalous result induced further inquiry, which not proving satisfactory, it was determined to submit the question to experiment, in order to prove how far these doubts were entitled to credit

To attain the objects of the experiments in a satisfactory manner, it was necessary that the apparatus for conducting them should be of great strength and large dimensions. For this purpose a cast-iron cylinder C, 8 feet long, 28 inches in diameter, and 2 inches thick of metal, was prepared for the

reception of the tubes to be experimented upon. A small pipe, a, was connected with a force pump, and by means of this, water was injected into the cylinder and the requisite pressure obtained. A second pipe, b, communicated with two steam pressure gauges, by which the force required for collapse was registered; and the indications of these were checked by a small and accurately fitted safety valve d. The large cylinder was fitted at top

and bottom with heavy ribbed covers, screwed to strong flanges on the cylinder, calculated to sustain great pressure. The tube to be experimented upon was fixed in the position shown at D, having cast-iron ends riveted and soldered to it to render it perfectly water-tight. The small tube m, communicating with the interior of the tube D, was for the purpose of allowing the escape of the contained air at the moment of collapse. The whole of the experiments were effected by means of the hydraulic pump, by which water was forced into the cylinder C; and the air, driven in a compressed state into the upper part, became highly elastic as the pressure was progressively increased until rupture took place. At very high pressures, the air in the cylinder C was permitted to escape, and collapse effected by water pressure only.

The tubes upon which the experiments were made varied from 18 inches to 60 inches in length; from 4 inches to 183 inches in diameter, and from 043 to 25 inch in thickness of metal. They were composed of plates of riveted sheet-iron, and the thinnest were carefully brazed at the joints to make them tight and prevent the entrance of water under pressure.

The results of the experiments may be stated under three heads: strength as affected by length, as affected by diameter, and as affected by thickness of metal.

I. Strength as affected by Length.-The results under this head are singularly interesting and conclusive. Within the limits of from 1.5 foot to about 10 feet in length, it is found that the strength of tubes similar in other respects, and supported at the ends by rigid rings, varies inversely as the length.

Thus, taking the four-inch tubes of different lengths, we have the following mean results derived from experiment:

(1.) Resistance of four-inch Tubes to Collapse.

Length.

ins.

19

60

40

Collapsing Pressure.

lbs. per sq. in.

137

43

65

The remarkable differences which exist in the resisting powers of the above similar tubes will be at once apparent. Assuming the experiment upon the tube 60 inches long to be correct, we may easily calculate the strength of the other 19- and 40-inch tubes, by the above-stated law of inverse proportion.

Thus, for the 40-inch tube, we have 40: 60 :: 43: x=64 lbs. And for the 19-inch tube,

19: 60: 43: x=135 lbs.,

where the calculated differ from the experimental results by 7th in one case, andths in the other.

(2.) Resistance of six-inch Tubes to Collapse.

Collapsing Pressure.
lbs. per sq.
in.

55

32

Here, from the data of experiment (1.), we may calculate the strength of a tube similar to that in experiment (2.)

59: 30: 55: x=28 lbs.,

where the calculated differs from the experimental result by th.

30:39:32: x=41 lbs.

differing from the result in (2.) by ths.

(4.) Resistance of ten-inch Tubes to Collapse.

Collapsing Pressure.

lbs. per sq. in.

19

33

or 14 lb. less than experiment (2.). In the same manner all the experiments might be taken and compared, and the law will be found to hold true in every case. The discrepancies are comparatively small, and, as they appear to follow no law, are evidently to be accounted for from defects in the construction of the tubes and difficulties in the mode of conducting the experiments, inseparable from such a mode of research.

II. Strength as affected by Diameter.-A precisely similar law is found to hold in relation to the diameter. Tubes similar in other respects vary in strength inversely as their diameters. Testing this law in the same manner as the last, we may at once place the calculated pressure beside that derived from experiment. Hence we have the following table:

(1.) Resistance to Collapse of Tubes five feet long.

The above variations are slight when compared with the resisting powers of the tubes. They were no doubt caused by the varying rigidity of the iron plates, or defects in the cylindrical form. Similarly we may take the results on tubes 30 inches long, and tabulate them in a similar manner.

(2.) Resistance to Collapse of Tubes 2 feet 6 inches in length.

As before, the variation between the results calculated by the law from