The real spore-bearing filaments of the Penicillium glaucum are only developed in air, for when the spores of that fungus are kept submerged in a liquid favourable for their growth, as in a saccharine solution, or the juice of the grape, they form an aquatic mycelium on the filaments of which cells are produced, and carbonic acid gas is given out. These cells increase by budding or division into chains of ferment, easily separated into single cells. The Penicillium glaucum is thus developed in all kinds of liquids, and in almost all kinds of conditions; even the peculiar knotty filaments observed in its submerged mycelium are not constant in different liquids. M. Hoffmann has observed with certainty the passage of Penicillium glaucum into Penicillium candidum, into a sulphur-coloured Penicillium, and, lastly, into the Coremium glaucum, so that this fungus is polymorphous, although the conditions under which the changes take place are unknown.[57] It appears that substances and liquids do not ferment spontaneously, for upon examining with a microscope the dust obtained by scraping the exterior of gooseberries, plums, vine leaves, potatoes, &c., M. Hoffmann found the short chains of the Torula, the necklace-like ferments of the Mucedines, and the chaplets of others. Some had already begun to germinate, and were developed readily when put into water. They had no doubt been carried by the wind from the dry refuse of fermented substances which are thrown away. [Illustration: Fig. 35. Torula Cerevisiæ, showing successive stages of cell multiplication.] The yeast of beer was at one time considered to consist entirely of the cells and chains of the Torula Cerevisiæ (fig. 35). This, which is one of the Coniomycetes of the order Torulacei, is in its early stage a colourless transparent globe capable of endless increase by budding. When in a liquid favourable to its growth, as the wort of malt, buds in the form of young cells spring from the walls of the globes; these soon become perfect and acquire buds also, so that in a few hours the parent cells develop themselves into rows of four or five globes, which remain in contact while the plant is growing, but separate if anything checks the fermentation. The full development of the plant only takes place when the fermentation is allowed to continue for some time, and then it is capable of producing a variety of forms, which show that it has other modes of reproduction besides budding.[58] In fact, when the fermentation is active in the upper parts of the liquid it appears in chaplets of from four to twelve articulations, or sometimes it ramifies into little branches.