INTERPRETATION 565 



between them and following them, so that sudden dehydration leads to 

 extreme distortion, (ii) The occurrence of a diffuse stage when, as 

 though in a resting stage, the chromosomes are lost to sight. Either of 

 these interruptions is fatal to the study of the succession of stages, 

 and both may occur in the same species (Chick, 1927, on Benacus). 

 These difficulties make the comparative method even more important 

 in studying meiosis. 



The application of the comparative method enables us to take a 

 more objective view of the nature of artefacts than is commonly done. 

 Artefact has been one of those words that " plainly force and overrule 

 the understanding." An artefact is an appearance that arises from 

 treatment. All living organisms are therefore liable to show artefacts 

 when they are killed, and they may, as Belar (1928, pp. 3-4) pointed 

 out, reveal " vital " artefacts while still showing signs of normal life, 

 e.g., the collapse of the chromosomes referred to above. These follow 

 merely from manipulation. But in ordinary treatment manipulation is 

 followed by fixation (coagulation and hardening) and staining (differen- 

 tial absorption of dyes). Both of these may lead to artefacts. The fact 

 that stains are apparently of three degrees of directness : (i) those 

 which stain cell-structure differentially in the first instance [e.g., Giemsa), 

 (ii) those which stain uniformly and are de-stained differentially 

 {e.g., gentian violet), (iii) those which are de-stained differentially after 

 mordanting [e.g., iron-haematoxylin), is probably not of great import- 

 ance, for the decisive differential reaction is in all cases a single one, and 

 judged by the results, under suitable conditions of fixation, it is of the 

 same degree of specificity. The differences in effect of treatment are 

 therefore almost entirely conditioned by differences in fixation to which 

 staining is secondary. 



No appearance of treated material is definitely free from artefact nor 

 is any appearance " pure " artefact. The question is therefore not so 

 much whether an appearance is an artefact but how significant the 

 artefact is. For this purpose comparison enables us to distinguish those 

 that are characteristic from those that are not, as follows : 



(i) The collapse of the chromosomes to one side of the nucleus is 

 characteristic of two stages of meiosis and is not found in the prophase 

 of the abortive meiosis of parthenogenetic aphides (Morgan, 19 12), 

 although it is found in the normal prophase. The reason for this has 

 been shown above, (ii) The chromomere structure of the early prophase 

 chromosome is a characteristic artefact because homologous chromo- 

 somes in the same nucleus show similar chromomeres. (iii) The spiral 

 structure seen after squeezing living chromosomes is characteristic, but 

 the fractured structure seen when they are squeezed after fixation is 

 not. (iv) The fibrillar structure of the shrunk spindle is characteristic 

 and is derived from a special arrangement of differently hydrated 



