For a thoroughly soft deep bog, however, it is most difficult to make any accurate calculation as to the amount of embankment filling which 71 will be necessary to form a permanent foundation for the line; and the construction of a high heavy embankment across such a place is one of those undertakings which every engineer is most anxious to avoid. A large quantity of material may be tipped into the bog, and seem to stand fairly well for a time, and then suddenly disappear altogether. More material has to be brought forward, and will most likely disappear in a similar manner. The filling material being heavier than the bog on to which it is thrown, falls through, and displacing the soft semi-liquid matter, continues to sink down lower and lower until it is stopped by a harder stratum underneath. In a measure the operation somewhat resembles the tipping of earth into a lake; the material will go down until it meets with a solid bottom, and in going down it assumes its own natural slope, and forms for itself a width of base corresponding to its height. It will be readily understood what an enormous amount of filling material will be swallowed up in following out such a process. On a very soft bog, say 20 feet in depth, over which an embankment 20 feet high has to be formed, the extent of the actual earthwork filling will very probably closely approach the outline shown in Fig. 63. The upper portion, ABCD, representing the embankment proper, will contain about 133 cube yards to the yard forward, whereas the lower portion, CDEF, which has displaced the soft boggy matter, will contain about 266 cube yards to the yard forward, or, in other words, the filling which is out of sight will be double the filling which is in view above the section ground line. Apart from the large amount of filling consumed in forming this semi-artificial island, the progress of the work itself is very perplexing. A long length of the bank may have been raised again, once or twice, to the proper height, and may have carried rails and earth-waggons for some weeks, and then sink all at once several feet. The sinking, too, may not be uniform, but may produce fissures, depressions, and separation of the earthwork which will necessitate much care when bringing forward fresh filling material. The bog may not be of the same consistency throughout, there may be some layers of harder material, such as imbedded trunks of trees, and these may sustain the filling for a time, and then yield under the increasing weight of the superincumbent mass. Even when the embankment is finished throughout, and shows no sign of sinking, it should be very carefully watched for a long time for any indication of further 72 movement.