THE. MICROBIOLOGY OF ALCOHOL PRODUCTS 517 



of protein matters that would serve as food for injurious bacteria. - Centrifugal 

 machines have been devised to hasten the process of defecation, but their work is 

 imperfect. 



Sterilization by heat has been tried for the same purpose but with indifferent 

 success. High heating caramelizes part of the sugar and oxidizes the must, thus 

 injuring the flavor. Discontinuous heating at lower temperatures in an atmospihere 

 of carbon dioxide is preferable but troublesome and expensive. All methods have the 

 defect of extracting undesirable substances from the solid matters which are heated ^ 

 with the must. 



Chemical sterilization is still less practicable. No substance could be used for 

 this purpose except sulphur dioxide; this used in sufficient quantities would seriously 

 injure the flavor of the wine. The effect would be totally different from that of the 

 smaE quantities used in defecation. 



All the methods discussed have for their object the diminution or elimination of 

 microorganisms of all kinds. With the injurious forms the true yeast is also re- 

 moved. The more perfect these methods, the more necessary it is to add wine 

 yeast. Without this addition, in fact, all these precautions may result in harm, for 

 the wine yeast, being present in much smaller numbers than many of the injurious 

 forms, may be completely removed while enough of other forms are left to spoil the 

 wine. 



A "starter" of some kind is therefore necessary with defecated must 

 and useful in all other cases. 



A Starter. — One method of producing a starter is to gather a suitable quantity 

 of the cleanest and soundest ripe grapes in the vineyard, crush them carefully 

 and allow them to undergo spontaneous fermentation. Perfectly ripe grapes 

 should be selected and the fermentation allowed to proceed until at least 10 

 per cent of alcohol is produced. If imperfectly ripened grapes are used or the 

 starter used too soon, the principal yeast present will be 5. apiculatus. Toward 

 the end of the fermentation, 5. ellipsoideus predominates. From 4 to 12 1. (i to 

 3 gallons) of this starter should be used for each 400 1. (100 gallons) of grapes or 

 must to be fermented. Too much starter should not be used in hot weather or 

 with warm grapes, otherwise it may be impossible to control the temperature. 



This starter is used only for the first vat or cask. Those following are started 

 from the first fermentations, care being taken always to use the must only from a 

 tank at the proper stage of fermentation and to avoid all tanks that show any 

 defect. 



An improvement on a natural starter of this kind is a pure culture of tested yeast. 

 Such yeasts are being used extensively in most wine-making regions, usually with 

 excellent results. The methods of use would require too much space to describe here, 

 but they are simple and such as could easily be devised by anyone with some knowl- 

 edge of microbiological technic. They do not aim at obtaining an absolutely pure 

 fermentation, which is unnecessary, but endeavor to have an overwhelming pro- 

 portion of a thoroughly tested and suitable yeast which wiU rapidly and perfectly 

 attenuate the wine before the few injurious microorganisms present have time to 

 do any harm. 



