Common sense, consideration, and judgment are all required in distinguishing what parts of girders and the like, and what members of trusses and other framed structures, are in compression, and what parts in tension.

With regard to the first-named - i.e., girders - the illustrations hereinafter given should clearly impress this distinction on the student's mind, the forms, of course, being exaggerated for the purpose.

Fig. 937 represents a girder supported only at each end, from which it will be seen that the whole of the top flange is in compression, and the bottom flange is entirely in tension.

Fig. 937.

Fig. 938 represents the same girder fixed at both ends - showing part of each flange in compression and part in tension - the top flange being in compression in the centre, "under the load," and in tension over the ends and at the supports, while the bottom one is exactly the reverse.

Fig. 939 is a combination of Figs. 917 and 938 - i.e., a girder having one end fixed and the other -supported, showing the different results; while Fig. 946 illustrates a continuous girder, having several supports, over which the top flanges, as will be seen, are in tension and the bottom ones in compression, the central portions being as usual, and as previously shown.

The best practical illustration of the above is a piece of indiarubber of sufficient length, subjected to similar conditions, etc., when it will easily be seen which part is in compression and which in tension.

Cantilevers, of course, must be fixed to hold themselves up, much more under a toad in addition; and when they are a simple projection, similar to half a girder, as Fig. 941, it will be seen that they are subjected to the strains of compression and tension, resulting from loads at their end, as indicated.

When they are framed it all depends on their form. Here the principle of considering the possibility of the substitution of rope solves all difficulties; as in Fig. 942, it will be apparent that a rope would do for the member A, which is therefore in tension; but not for B, which is consequently in compression. On referring to Fig. 943, it will be at once seen that exactly the opposite is the case, A being in compression and B in tension, the dotted lines showing the effect of an excessive load.

Turning to the consideration of roof trusses, we find the application of the rope principle more useful than ever; but it should first be pointed out that the weights on roofs are, as it were, collected together and located at various points, proportionately, as Fig. 944, from which it will be seen the eaves joint A take, 1/8 part on each side, while the ridge C, and purlin joints B, each take 2/8 or 1/4 of the load.

To discriminate between the members in compression and tension, we find, on reflection, that a rope would certainly not act as an efficient substitute for the rafters or struts, as shown by thick lines, but that it might be used for the tie and suspension rods shown by thin lines (see Fig. 944). Therefore the latter are in tension, and the former in compres sion or transverse strain.

Figs. 945 and 946 are further illustrations of trusses, the latter being the ordinary wooden king post; Fig. 947 represents a trussed beam, a principle often employed to strengthen thin boards to carry flower pots, the thin line members being usually of string.

Fig. 945.

Having ascertained which member or part is in tension, and which in compression, we know that wrought iron has the greatest power to resist strains tending to elongate it or tear it asunder; therefore we can do no better than use it for members in tension, and, generally speaking, it is most convenient in round or bar form, though it can be rolled into T's for struts and rafters. As regards parts in compression, cast iron, on account of its great power to resist compression, would be the strongest and best, were it not for its fragile, unweldable, and inconvenient nature, for the putting together of the joints. As a consequence, wrought iron has superseded cast iron for rafters and struts which are made of T shape, the form most suitable for long lengths, where bars or rods would bend, unless built up to make them the more rigid, as hereinbefore shown.

Fig. 946.

Fig. 947.

Joists, rafters, lintels, etc., subject to transverse strain (though rafters have been taken above as in compression, but only in contra-distinction to tension), are strongest when of rectangular form and of a good depth, the strength lying in the depth, and not in the breadth.