The anonymous _Mémoire_ of 1718 gives, with an amount of preliminary flourish which would imply a doubt as to the accuracy of the statement made, the secret mode said to have been employed by Dom Perignon to improve his wine, and to have been confided by him a few days before his death to 'a person worthy enough of belief,' by whom it was in turn communicated to the writer. According to this, a pound of sugar-candy was dissolved in a _chopine_ of wine, to which was then added five or six stoned peaches, four sous' worth of powdered cinnamon, a grated nutmeg, and a _demi septier_ of burnt brandy; and the whole, after being well mixed, was strained through fine linen into a _pièce_ of wine immediately after fermentation had ceased, with the result of imparting to it a dainty and delicate flavour. Dom Grossart, however, in his letter to M. Dherbès, distinctly declares that 'we never did put sugar into our wine.'[110] This _collature_, in which peaches play a part, was probably made use of by some wine-growers; and the peach-like flavour extolled by St. Evremond in the wine of Ay may have been due to it, or to the practice then and long afterwards followed of putting peach-leaves in the hot water with which the barrels were washed out, under the idea that this improved the flavour of the wine.[111] Opinions were widely divided as to the cause of the effervescence in the wines of Hautvillers, for the connection between sugar and fermentation was then undreamt of, although Van Helmont had recognised the existence of carbonic acid gas in fermenting wine as early as 1624. Some thought it due to the addition of drugs, and sought to obtain it by putting not only alum and spirits of wine, but positive nastinesses, into their wine.[112] Others ascribed it to the greenness of the wine, because most of that which effervesced was extremely raw; and others again believed that it was influenced by the age of the moon at the epoch of bottling. Experience undoubtedly showed that wine bottled between the vintage and the month of May was certain to effervesce, and that no time was more favourable for this operation than the end of the second quarter of the moon of March. Nevertheless, as the wines, especially those of the Mountain of Reims, were not usually matured at this epoch, it was recommended, in order to secure a ripe and exquisite sparkling wine, to defer the bottling until the ascent of the sap in the vine between the tenth and fourteenth day of the moon of August; whereas, to insure a _non mousseux_ wine, the bottling ought to take place in October or November.[113]