The Britons set fire to the Roman Camp during Cæsar’s second invasion, 54 B.C., by discharging hot balls of clay among the tents.[566] At the attack on Placentia, A.D. 69, igneous missiles were employed (_glandes et missilem ignem_), and probably destroyed the amphitheatre.[567] As before mentioned, hot shot (for cannon) were invented by the Polish king, Stephen Bathory, in 1579.[568] Their greatest triumph was the destruction of d’Arçon’s floating batteries and a great part of the Spanish fleet at Gibraltar, 13th September 1782. _Incendiary Fireballs._ The gunners of old encountered great difficulties in their endeavours to introduce igneous projectiles. Their use in the early guns was not absolutely impossible, but it would have been fruitless; for to prove effective an igneous projectile, whether incendiary or explosive, must contain a considerable mass of combustible matter, and this condition could not be fulfilled with guns of very small calibre. When the calibre had greatly increased, during the last quarter of the fourteenth century, any attempt to employ such igneous projectiles as were in use with the machines must have ended in failure. The action of the machines was similar to that of a sling, and the shells (or envelopes) of their incendiary missiles were made just strong enough to resist the pressure to which they were subjected on discharge, although not strong enough to bear the shock of impact with the object they struck. This broke them up and scattered their blazing contents about. Such projectiles were evidently unfit for use in cannon; for the explosion of the charge would inevitably break them up in the bore, and their viscous contents would travel but a very short way. Owing to these difficulties the machines held their ground to the middle of the fifteenth century, if not longer, and the igneous projectiles ultimately constructed for cannon were developments of the hand-grenade. In Fig. 31 of the plate from the MS. of Kyeser’s “Bellifortis,” 1405, given by Herr von Romocki (i. 169), we are shown a projectile which unquestionably belongs to the same family as the _tonneau_ which terrified Joinville and his companions;[569] but this barrel could have only been discharged from a machine. Whether Figs. 26 and 28 of the same plate were thrown by hand or machine depended on their size, which we do not know. From their construction, with a mere covering of cloth or cordage, we may safely conclude that they were not gun-projectiles.