would be exceedingly increased by the consequent additional pressure, while the process of injection would not be in the smallest degree facilitated by it. Trap-veins would thus be produced, affording indubitable evidence of injection.

Again, the hypothesis we have made above of the entire absence of adhesion between two contiguous beds, though it may in some cases be true for limited spaces, cannot be uniformly so in cases in nature for spaces of considerable extent. Now in those instances, in which the force of adhesion between the two beds, bears any kind of ratio to that which holds together the component particles of the uplifted portion, an enormous force will be required to overcome this adhesion. And how are we to conceive such a force applied without producing the smallest rupture in the lower surface of the uplifted mass? If there be no adhesion between the beds, no considerable horizontal tension will be produced in this mass; but if the adhesion be considerable, such a tension will be produced, proportional to the increased force of injection called into action. Under these circumstances the smallest break or crevice will be torn open, the fluid matter will enter it, and acting on its vertical sides with an enormous pressure, and with the mechanical advantage of a wedge, will add immensely to the tendency of the horizontal tension to produce a vertical fissure.

78. It may perhaps be thought that the difficulty of conceiving the process of horizontal injection of considerable extent, without the production of vertical veins, may be obviated by supposing the fluid

Α

matter injected from many points simultaneously between the same two horizontal beds. But this hypothesis appears extremely improbable,

unless it be also assumed that the want of cohesion between these beds is co-extensive with the injected bed, an assumption, which as I have before remarked, must probably be in general considered as totally inadmissible. The probable consequence of simultaneous injection from different fissures, (supposing the injected matter not in too great quantity), would be the formation of partial and unconnected beds as represented in the annexed diagram.

For these reasons then we cannot hesitate, I think, to conclude, when we consider the general structure of stratified masses, that the absence of numerous trap-veins and dykes, having their origin in the upper surface of a horizontal bed of trap, with the want also of very frequent indications of violent mechanical action in the lower portion of the superincumbent mass, affords indubitable proof of the fact of such horizontal bed having been ejected over the exterior surface existing at the time of its eruption.

79. The existence of a single vein or dyke such as above described, in rocks incumbent on a horizontal bed of trap, is clearly an indubitable proof of injection; but it must not therefore be concluded, that every trap-vein or dyke in the superincumbent strata affords this unequivocal testimony, since it is manifest that such a vein or dyke might possibly be produced by injection, subsequently to the formation of the horizontal bed, which it may have traversed exactly in the same manner as any other stratum*. The decisive character of the evidence of injection afforded by a vein, consists in its originating in the upper surface of the injected bed. We may also remark, that indications of mechanical action on the beds beneath a bed of trap will not necessarily afford conclusive testimony as to the fact of injection, because such appearances might be produced, to a certain extent, by the force of an ejected, as well as of an injected bed. It is in the superincumbent beds that we must seek for the evidence in question.

80. It is not my object to enter into any detailed comparison between observed facts, and these theoretical deductions, but I think it necessary

Many instances are given by M'Culloch of veins of trap existing in trap. See "Description of the Western Islands."

to allude to some cases of injection described by M'Culloch in the Western Islands, in which the injected beds assume for considerable distances the appearance of being regularly interstratified, thus seeming, it might be thought, to offer exceptions to the rule I have deduced from theoretical considerations. Four or five only of these exceptions I think have been expressly mentioned by that author. Those on the coast of Trotternish in the Isle of Sky, which appear to be the most striking, are described as follows:

"In one case, which occurs not far from Holme, there is a bed extending for a great way, surmounted by a parallel series of the secondary strata in contact with it; but on a narrow inspection, innumerable veins are seen branching into the strata in every possible direction, illustrating in a very perfect manner the origin of at least one order of veins. In a second case, three beds of trap can be traced in a parallel direction for a considerable space, separated by the regular strata, when suddenly the whole unite into one mass. Had not this occurrence at length betrayed the true nature of these beds, there would have been no hesitation, from a limited observation, in describing them as unquestionable instances of alternation. In the last case which I shall mention, one regular bed of trap may be traced for more than a mile, lying in a parallel and undisturbed continuity between the secondary rocks. On a sudden, however, it bends downwards so as to pass through the strata immediately in contact, and then continues to hold its regular course for a space equally great, with a thickness and parallelism as unaltered as before*."

The first of these instances presents in its branching veins, exactly the phenomena which, I have been contending, must necessarily attend any extensive horizontal injection of a fluid mass. The others seem to indicate the possibility of this injection without such phenomena, for at least the extent of a mile. Nor am I disposed to doubt this possibility, though I should in general consider a horizontal injection of that extent without ramifying veins, as extremely improbable, and especially if the injected bed were not a very thin one. In fact, however, there Description of the Western Islands of Scotland, Vol. 1. p. 382.

appears no reason to conclude that such has been the case. The whole of Trotternish is described as consisting of an enormous overlying mass of trap, which appears to have risen in numberless places through the stratified rocks on which it reposes. It extends (if I understand the description rightly) quite to the coast, so that scarcely any stratified rocks are visible, except in the vertical section formed by the steep cliffs along the beach, and in which the appearances above described are observed. Hence it is probable that these horizontal beds are connected with vertical masses of trap, at distances from the visible sections of them, small in comparison with their apparent range along the cliffs, and consequently it is very possible that the extent of horizontal injection may have been much less than at first sight it appears to have been.

The same observation will apply to the other phenomena of the same kind as described by the author just quoted; and so far from offering any thing opposed to the theoretical views I have been explaining, they may, I think, be considered, when taken in conjunction with numberless cases of vertical dykes and veins, as strongly corroborative of them; since the comparatively insignificant number of these injected horizontal beds, clearly proves them to offer only so many exceptions to the very general rule of verticality in trap-veins, so frequently recognized by M'Culloch himself.

81. In speaking of horizontal injection, I have not yet alluded to the consequences of imperfect fluidity in the injected matter. If we may be allowed to judge of the degree of this fluidity from the analogy which the injected matter may be presumed to have borne to modern lava in its eruption, we may conclude it to have fallen considerably short of that of perfect fluidity. Consequently the lateral pressure communicated by the fluid, would never be equal to the direct pressure impressed upon it, and this, it is evident, would increase the difficulty of horizontal injection in the cases which I have already considered. The most important consideration, however is, I conceive, that this property of imperfect fluidity, would thus impose a limit to the possible extent of lateral injection, supposing the injected matter not to form a bed lying in one plane, but to form an irregular surface, such

th

that the following diagram may represent a vertical section of it. For suppose the fluid capable of transmitting the () of a force impressed upon it, in a direction perpendicular to that of the impressed force; then if the pressure be transmitted along a broken line consisting of straight lines at right angles to each other, it is clear that the force transmitted along the first straight portion (supposed horizontal), will be being the impressed force and acting vertically. Along the second portion

2

m

n

of the broken line the transmitted force will be (). p, and generally

B

'm

along the 7th portion it will be (). p. If the different portions of the

n

broken line be not at right angles to each other, or instead of being straight be curved, the diminution of the transmitted pressure must still be calculated on the same principle. It is important, however, to observe that the thickness of the injected bed would probably influence this diminution very materially, as may be illustrated by the following figure. If the section of the bed be represented by the space between

a

b

b...

В

the lines ab and cd, a straight line may be drawn in it from one