The Chemistry of Cookery

be observed that a special fire directly under the roaster is here designed, and that this fire is enclosed in brickwork. This is a general feature of Rumford’s arrangements. The economy of the whole device will be understood by the fact that in a test experiment at the Foundling Institution of London, he roasted 112 lbs. of beef with a consumption of only 22 lbs. of coal (three pennyworth, at 25_s._ per ton). Rumford tells us that ‘when these roasters were first proposed, and before their merit was established, many doubts were entertained respecting the taste of the food prepared in them,’ but that, after many practical trials, it was proved that ‘meat of every kind, without any exception, roasted in a roaster, is _better tasted, higher flavoured, and much more juicy and delicate_ than when roasted on a spit before an open fire.’ These italics are in the original, and the testimony of competent judges is quoted. I must describe one experiment in detail. Two legs of mutton from the same carcass made equal in weight before cooking were roasted, one before the fire and the other in a roaster. When cooked, both were weighed, and the joint roasted in the roaster proved to be heavier than the other by 6 per cent. They were brought upon table at the same time, ‘and a large and perfectly unprejudiced company was assembled to eat them.’ Both were found good, but a decided preference given to that cooked in the roaster; ‘it was much more juicy, and was thought better tasted.’ Both were fairly eaten up, nothing remaining of either that was eatable, and the fragments collected. ‘Of the leg of mutton which had been roasted in the roaster, hardly anything visible remained, excepting the bare bone, while a considerable heap was formed of scraps not eatable which remained of that roasted on a spit.’ This was an eloquent experiment; the gain of 6 per cent. tells of juices retained with consequent gain of flavour, tenderness, and digestibility, and the subsequent testimony of the scraps describes the difference in the condition