32 



Methods and Techniques 



The significance, however, of the behavior 

 of a part removed from its surroimdings is 

 lost except in comparison with its behavior 

 in those surroundings, as Roux already knew. 

 In most cases, except where natural pig- 

 ments are present, the observation of the 

 egg as a whole sheds all too little light on 

 the separate activities of its individual parts, 

 and Roux himself attempted to circumvent 

 this difficulty by inventing a crude marking 

 experiment and by pricking his eggs to pro- 

 duce extra-ovates which might serve as mark- 

 ers. This experiment is open to the obvious 

 criticism that it may alter the status of the 

 part whose normal behavior it purports to 

 elucidate, and it was to obviate this that the 

 technique of local vital staining was devel- 

 oped. It has reached its highest perfection 

 as developed for application to the amphibian 

 egg by Vogt ('25), where its success depends 

 on the fact that inclusions of the egg adsorb 

 the stain from the carrier more rapidly than 

 it diffuses into the solution, and a mark of the 

 utmost sharpness of outline is therefore 

 achieved and maintained. 



The data obtained by the local vital stain- 

 ing method are indispensable for the inter- 

 pretation of experiments in which the ac- 

 tivities of particular parts are to be studied 

 by other means, and the dyes currently used 

 (Nile blue sulfate, neutral red and Bismarck 

 brown) are sufficiently nontoxic that the data 

 derived from their use are thoroughly valid. 

 There are conditions, however, under which 

 the method has strong limitations. In some 

 cases the dyes may be transformed to leuko- 

 bases within the cells. In the case of embryos 

 whose cells lack inclusions with special affin- 

 ity for the dyes, for instance young stages of 

 chick and teleost, the stains are far more dif- 

 fuse and ephemeral than in the amphibian, 

 and the results of their use may prove unreli- 

 able when checked against results obtained by 

 other methods. The newer method of follow- 

 ing morphogenetic movements by the appli- 

 cation to the cells of carbon particles (Spratt, 

 '46) has, for instance, produced results for 

 the chick which are incompatible with those 

 previously derived for the same form by 

 vital staining (Pasteels, '37). The use of car- 

 bon particles holds great promise for the 

 future, but the introduction of macroscopic 

 particles within the cells raises certain dan- 

 gers for the interpretation of what are sup- 

 posed to be unhampered movements; and 

 when the particles are applied to the outer 

 surface of the cell there may always remain 

 some doubt as to whether they may have 

 shifted in position. 



So far as the localization and retention of 

 a marker is concerned, the least reproach- 

 able method of distinguishing one group 

 of cells from another remains the observation 

 of forms in which natural pigments occur. It 

 may be remembered in this connection that 

 the method of heteroplastic grafting was at 

 its earliest inception used to trace the migra- 

 tion of elements distinguished by natural 

 pigment (Harrison, '03; see Harrison, '35, for 

 later uses of heteroplastic grafting) . Grafting, 

 however, as an operative method, introduces 

 new sources of error not inherent in the 

 methods of marking cells in an unoperated 

 embryo. Unfortunately there is still no equiv- 

 alent of a Geiger counter to report on the 

 migrations of cells in the embryo. The data, 

 however, which have been accumulated by 

 the present techniques as applied by cautious 

 investigators are adequate to serve as a 

 frame of reference against which studies of 

 the parts may be judged. 



Since killing a cell is probably the easiest 

 thing that an embryologist can do, it was per- 

 haps inevitable that the study of the behavior 

 of parts of an embryo should have first been 

 examined by defect methods, and it is ap- 

 propriate next to mention some of the ways 

 in which defects have been produced. Experi- 

 ments involving the removal or the supposed 

 inactivation of cells or their parts have been 

 carried out by mechanical, chemical, thermal 

 and electrical methods and by combinations 

 of them, and by the use of various sorts of 

 radiations. 



A primary and insidious source of error 

 common to all these methods is that in ap- 

 plying them the investigator may alter more 

 factors than he knows. The classical experi- 

 ment of Roux (1888) was designed by killing 

 a blastomere to eliminate its influence on its 

 neighbor: the fact that its corpse exerted 

 mechanical influence of moment could be ap- 

 preciated only after McClendon ('10) com- 

 pletely removed a blastomere to demonstrate 

 a different accomplishment by the remaining 

 cell than had been achieved in Roux' experi- 

 ment. Comparable soiu-ces of error may lie 

 hidden in many if not all of the defect ex- 

 periments subsequently performed. 



In dealing with the deletion or inactiva- 

 tion of components of cells, the perils of in- 

 terpretation may be as great. Boveri ('18) 

 long ago recognized as an inevitable source 

 of error in experiments designed to exclude 

 nuclear influence that nuclear residues might 

 remain undetected in the cytoplasm, which 

 in any case has necessarily been produced 

 under the influence of the nucleus. The cen- 



