There is no longer much doubt but that both acid-forming and casein-digesting species can take part in the production of proper flavors as well as desirable aromas. The researches of Conn,[159] who has studied this question most exhaustively, indicate that both of these types of decomposition participate in the production of flavor and aroma. He has shown that both flavor and aroma production are independent of acid; that many good flavor-producing forms belong to that class which renders milk alkaline, or do not change the reaction at all. Some of these species liquefied gelatin and would therefore belong to the casein-dissolving class. Those species that produced bad flavors are also included in both fermentative types. Conn has found a number of organisms that are favorable flavor-producers; in fact they were much more numerous than desirable aroma-yielding species. None of the favorable aroma forms according to his investigations were lactic-acid species,--a view which is also shared by Weigmann.[160] McDonnell[161] has found that the production of aroma in certain cases varies at different temperatures, the most pronounced being evolved near the optimum growing temperature, which, as a general rule, is too high for cream ripening. The majority of bacteria in ripening cream do not seem to exert any marked influence in butter. A considerable number of species are positively beneficial, inasmuch as they produce a good flavor or aroma. A more limited number are concerned in the production of undesirable ripening changes. This condition being true, it may seem strange that butter is as good as it is, because so frequently the requisite care is not given to the development of proper ripening. In all probability the chief reason why this is so is that those bacteria that find milk and cream pre-eminently suited to their development, e. g. the lactic-acid class, are either neutral or beneficial in their effect on butter. ~Use of starters.~ Experience has amply demonstrated that it is possible to control the nature of the fermentative changes that occur in ripening cream to such an extent as to materially improve the quality of the butter. This is frequently done by the addition of a "starter." While starters have been employed for many years for the purpose mentioned, it is only recently that their nature has been understood. A starter may be selected from widely divergent sources, but in all cases it is sure to contain a large number of bacteria, and the presumption is that they are of such a nature as to produce desirable fermentative changes in the cream.