C. F. Robinow 



419 



tinic structures by 45-90 min. incubation (i.e. during 

 the lag phase) on a fresh nutrient medium are studied. 

 The results of such observations show that all but a small 

 proportion of the bacteria undergo the growth changes 

 described in the preceding sections. 



(5) Atypical forma. To complete this account of the 

 chromatinic structures, three abnormal types of cells 

 must be mentioned which occur in cultures of Bact. coll, 

 Proteus vulgaris and many other species (Text-fig. 3). 

 Of these, the most interesting are bacteria of normal 

 length which apparently are not divided into two cells; 

 they have a faintly stained dumbbell body at each end 

 and a more deeply stained pair of these bodies near the 

 middle. 



The second atypical form may be short or several 

 times the normal length and contains a variable number 

 of chromatinic dumbbells compacted into a central 

 deeply staining mass. The third and fourth forms are long 

 filaments with many discrete but irregularly spaced 

 chromatinic bodies. Though always rare, the last two 

 types are more common in young cultures derived from 

 very old (e.g. 4 weeks) material, than in subcultures from 

 slants incubated for only 18 hr. Both have been ac- 

 curately described and illustrated by Neumann, whose 

 inability to determine their function I unfortunately 



Speculation about the significance of all these enig- 

 matic cells such as their connexion with autogamic pro- 

 cesses, is bound to be unprofitable until it is possible to 

 isolate them from normal living material and follow 

 their development separately. 



Text-fig. 3. Diagram of certain peculiar types of cells, 

 small numbers of which are regularly found in cultures 

 of Proteus vulgaris and Bact. coli. Tentatively arranged 

 in order of what may be their true genetical relation- 

 ship. At present any such connexion between the 

 different types is purely hypothetical. 



DISCUSSION 



The chromatinic structures described in this communi- 

 cation obviouslj correspond to those which J. Badian, 

 in a scries of papers since L930, has reported in the cells 

 of myxobacteria, actinomycetes and aerobic spore- 

 forming bacteria and which were also described in the 

 spores and vegetative forms of />'. mycoides and !>. mesen- 

 tericus by Robinow (1!U2). There is good reason to 



believe that SUCh _ st n let ill e- occur \ erv widelj among 



bacteria. 



The chromatinic dumbbell bodies are also identical 

 with the 'nucleoids' of Piekarski (1937 40). The differ- 

 ence between Piekarski's and my own description of the 

 nuclear structures is twofold and is due to differences 



in technique. According to Piekarski the nucleoids are 

 round bodies and each ceil has two, one near each end. 

 My observations make it clear that (1) the nucleoids in 

 Piekarski's preparations are optically unresolved dumb- 

 bell bodies, or, more often, configurations of these and 

 (2) that the 'cell' or ' Primaerform ' of his terminology is 

 actually a two-cell rod with a resting or dividing dumb- 

 bell body in the centre of each of its component cells. 

 In other words: the binucleate form which Piekarski 

 puts at the beginning of his sequence of developmental 

 stages is really composite and derived from small, uni- 

 nucleate coccoid or rod-shaped elements (Piekarski's 

 ' Sec undaer form') through growth and cell division. This 

 process is evident during the lag-phase and is less 

 common during the first few hours of a culture's active 

 growth, the stage to which Piekarski paid particular 

 attention. 



Neumann's (1941) observations on Bact. coli and 

 Proteusagree with those of Piekarski. Neumann observed 

 chromatinic dumbbell bodies but thought they were 

 stages in the division, hour-glass fashion, of a single 

 chromatinic granule identical with Piekarski's nucleoids. 

 Neumann was not aware of the multiple cell structure of 

 rod-shaped and filamentous bacteria, but apart from 

 these and other, less important, differences my observa- 

 tions on Proteus vulgaris and Bact. coli fully confirm 

 Neumann's detailed description of the distribution of 

 nuclear material in the various growth forms of these 

 organisms. 



There is some diversity of opinion about the mode of 

 division of the chromatinic bodies in different groups of 

 bacteria. This problem is important in connexion with 

 the possible function of these bodies in the transmission 

 of hereditary characters. 



In the rod-shaped bacteria, my own observations indi- 

 cate that the dumbbell bodies divide like chromosomes 

 by splitting longitudinally. The compact nuclei which 

 Beebe (1941) describes in the vegetative cells of My xo- 

 coccus xanthus n.sp. and which he believes divide amito- 

 tically by constriction, are more probably composed of 

 closely contiguous dumbbell bodies. The appearance of 

 a compact nucleus dividing by constriction is easily 

 simulated when, as described in the present communica- 

 tion, the chromatinic dumbbell bodies divide again 

 before they have completely separated after the previous 

 division, and especially if coloured strands persist be- 

 tween two pairs of closely contiguous bodies as they 

 move towards opposite poles (cf. PI. 5, figs. 1 e, 9o; 

 PI. 7, fig. 27 a). This interpretation of Beebe's results is 

 the more likely since he himself describes the nuclei of 

 cells migrating to the fruiting centres as breaking up into 

 'four irregularly shaped bodies or chromosomes'. An 

 alternation of amitotic and mitotic division such as this 

 author envisages seems hardly likely. 



Direct division of a round nuclear body by elongation 

 and constriction has recently been described by Knyasi 

 (1!U2) in a new species of yellow coccus. I have pre- 

 viously expressed the view (Robinow, 1942) that the 

 Feulgen positive granules which I noted in most of the 

 cells of two sarcinae were the ends of dumbbell bodies 

 w liieh divided by longitudinal splitting. I now find that 

 my photographs are more satisfactorily explained by 

 Knyasi's assumption that the granules divide directly, 

 the smaller cells containing one and the larger cells two. 



178 



