Crystals are not found in the luminous organs of animals with the exception of the fireflies. In these a layer of cells occurs (see Chapter IV) filled with minute crystals of one of the purine bodies (xanthin or uric acid). One might surmise that the light of the animal was a crystalloluminescence accompanying the formation of these crystals. It is easy to show, however, that the light comes not from the crystal layer but from another layer of cells containing large granules. It is also dependent on the presence of oxygen while crystalloluminescence takes place in the absence of oxygen. The crystal layer possibly serves as a reflector. Its significance will be discussed in a later chapter. [Illustration: FIG. 5.--Dubois's figures showing transformation of photogenic granules to crystals (_after Dubois_).] The light of luminous organisms is quite generally associated with granules. In one of the centipedes (_Orya barbarica_), which produces a luminous secretion, Dubois (1893) has described the transformation of these granules into crystals and at one time he supposed the light to be a crystalloluminescence. He later reversed this opinion and, certainly, examination of his drawings which are reproduced in Fig. 5 does not convince one of the actuality of crystal formation. The phenomenon of _lyoluminescence_, described by Wiedemann and Schmidt (1895) as a light accompanying the solution of colored (from exposure to cathode rays) crystals of Li, Na, or K chlorides, is probably due to a triboluminescence from stirring of the crystals during solution. CHEMILUMINESCENCE.--As the name implies, chemiluminescence is the production of light during a chemical reaction at low temperatures. This does not mean that the other types of luminescence are not connected with chemical reactions--using the word _reaction_ in a broad sense--for we have reason to believe that in some cases spectra are not characteristic of the element as such but are rather characteristic of a particular reaction in which the element takes part (dissociation into ions, changes from monovalent to bivalent condition, etc.) and that this is the reason one element may show various spectra under different conditions (Bancroft, 1913). The chemiluminescences are rather oxidation reactions involving the absorption of gaseous or dissolved oxygen and may be very easily distinguished from all the previously mentioned luminescences by this criterion. They should, perhaps, more properly be called _oxyluminescences_. The glow of phosphorus is the best known case, recognized since phosphorus was first prepared by Brandt in 1669. It is interesting to note that when first prepared phosphorus was regarded as a peculiarly persistent type of phosphor, _i.e._, a material akin to the impure alkaline earth sulphides.