in fire appear in the air." These letters are made on a frame covered with composition, and are consequently limited to a size to the internal diameter of the shell, that is, less than eight inches. It seems improbable that they could be distinguished satisfactorily at the height of a shell's trajectory, besides which the difficulties involved, as he himself explains, are very great, which no doubt explains the fact that this idea is now obsolete.
Under the heading "Double and Triple Balloons," this writer describes the method of placing shell of smaller size inside a larger. The bursting of the first shell lights the short-time fuse of the contained shell, which falls some distance and bursts. With the triple shell this action is repeated.
Jones (1765) divides shell into four kinds, namely, "illuminated balloons" filled with stars; "balloons of serpents," "balloons of reports, marrons and crackers," and "compound balloons." The last description is misleading, as the balloon is not compound but the contents are varied, as for example, the contents of one specified ten crackers of six reports, twenty golden rains, sixteen two-ounce cases charged half-inch with star composition and bounced, two ounces each of brilliant, blue, coloured tailed, large string and rolled stars. It is hard to believe that this writer had ever seen a shell fired in this manner, the result would have been mere confusion. The star compositions of that date were very rudimentary, the colours when seen from the distance of a bursting shell were indistinguishable.
One interesting detail in Jones's work is the classification of sizes. The smallest shell mentioned by him is the "Coehorn Balloon"; he does not give the size, but it is given in the "Military Encyclopedia" as 4-2/3 inches. This corresponds to the 4-1/2-inch of to-day. The name was apparently derived from
