1. The writer has followed up his earlier work on the dementia præcox group (1910) with a more systematic anatomoclinical study of 25 cases, having a view to (a) definite conclusions as to the structurality ("organic nature") of the disease, and (b) correlation of certain major symptom groups (delusions, catatonic symptom groups, auditory hallucinosis) with disease of particular brain regions.2. As to (a), the structuratity of dementia prœcox, the writer feels that the disease must be conceded to be in some sense structural, since at least 90 per cent of all cases examined (50 cases, data of 1910 and 1914) give evidence of general or focal brain atrophy or aplasia when examined post mortem, even without the use of the microscope.3. Moreover, with the use of the microscope, the problem of the normal-looking remainder can perhaps be solved, since the only two normal-looking brains7 in the 1914 series of 25 yielded abundant appearances of cell-destruction and satellitosis in the cerebral cortex, which had not yet had time to be registered in the gross (cases of three weeks' and two months' duration respectively).4. The method of anatomical analysis in the new series is a more systematic one than has been hitherto employed, involving careful gross description of the fresh brain; careful preservation (by suspension from basal vessels) in formaldehyde solution; systematic photography to scale of the superior, inferior (cerebellum removed), lateral, and mesial aspects before and after stripping the pia mater; study of all aspects of the brain as spread side by side in photographic form; further study of the preserved brains in the light of the photographic study; and eventual cytological or fiber studies of paired structures showing possible atrophy or aplasia.5. The neuropathologist making such a brain analysis shortly discovers that there is often more to be learned from the gross than from the microscopic appearances, since, of two gyri, the one measurably smaller than the other (and therefore probably agenesic, aplastic or atrophic), the microscopic appearances may often be hard to diagnosticate, as the normal-looking gyrus at the time of death may be just undergoing a satellitosis actually indicating more disease than its shrunken fellow.6. Nevertheless, the gross analysis gives one perfectly convincing evidence of some kind of lesions, leaving to other methods of study the decision as to the congenital or acquired nature of these lesions. Some 14 of the 25 cases may be regarded as in some sense maldevelopmental, so as to arouse the suspicion that the acquired atrophy was grafted on top of a congenital agenesia or aplasia; but, in the opinion of the writer, aplasia is indicated rather than agenesia: the potential victim of dementia præcox is probably born with the normal stock of brain cells, although their arrangement and development are at times early interfered with.7. The atrophies and aplasias, when focal, show a tendency to occur in the left cerebral hemisphere. The coarse atrophy is usually of only moderate degree, and often does not appreciably alter the brain weight, at least outside the limits of expected variation. In fact, the heart, the liver, the kidneys, and the spleen tend to show greater loss in weight than does the brain.8. More remarkable than the atrophy and aplasia of the cortex is the high proportion of cases of internal hydrocephalus (at least nine cases) uncovered by the systematic photographic study of frontal sections.9. There is no evidence that this internal hydrocephalus is due to generalized brain atrophy. It is possible that it begins more posteriorly. It is probable that it does not mechanically so much affect the frontal lobes. It is associated with cases of long duration, although not with all cases of long duration, and was never found in cases of brief duration. Clinically, the hydrocephalic cases are uncommonly catatonic, and the cases of marked generalized hydrocephalus were as a rule victims of hallucinations. Delusions, except fantastic delusions, were not prominent in this group. The clinical courses of these hydrocephalic cases were more than usually active and mutable, and were often interrupted by remissions.10. The hydrocephalic brains were not in other respects particularly open to the suspicion of congenital disease; and, without adequate proofs, the writer is inclined to consider the hydrocephalus to be often an acquired hydrocephalus.11. An ardent supporter of congenital features might claim that 19 of the 25 brains showed some sort of maldevelopmental defect; one impartial witness thought that 14 showed such; and even if all nine cases of hydrocephalus be taken as acquired, we remain with II cases bearing pretty certain evidence of maldevelopmental defect. On the other hand, all but six cases showed signs of acquired lesion, and these six showed various microscopic changes of doubtful meaning, but certainly acquired.12. One remains with the general impression that gross alterations are almost constant and microscopic changes absolutely constant, and that the high proportion of gross appearances suggesting aplasia means that structural (visible or invisible) changes of a maldevelopmental nature lie at the bottom of the disease process. But this suspicion of underlying maldevelopment is only a suspicion, although a strong one, and the first factor for the theory of pathogenesis to explain is the gross and microscopic changes as they present themselves in the full-fledged case.13. Aside from left-sidedness of lesions and internal hydrocephalus, very striking is the preference of these changes to occupy the association-centers of Flechsig. For this there is probably good à priori reason in the structure, late evolutionary development, and consequent relatively high lability of these regions. The interest of these findings is still greater in the functional connection (see below).14. In concluding this summary of the anatomical side of the study, the writer cannot forbear adding that he supposes many neurologists, hearing of "lesions," will at once imagine extirpatory lesions of a Swiss-cheese appearance or areas like those of tuberous sclerosis. At the risk of being charged with naïveté, the writer would again here insist that the lesions described, though never beyond the range of a skilful anatomist, are of a mild atrophic nature or in the nature of aplasias, requiring care and deliberation in their description and explanation, and often hard to grasp except where photographs of all sides of the brain may be compared at once and reference then made to the brains themselves. These lesions do not effect globar lacunle in the cortical neuronic systems, but they are of a more finely selective character. Under the microscope it may be difficult to say, without elaborate micrometry, that one area is worse off than another; but convincing evidence of the gross convolutional extent of the process is got by the naked eye and by the finger.15. The writer regards this work as putting the burden of proof on those who claim the essential functionality of dementia præcox, and is at some pains to couch objections to one formulation of these changes as "incidental," and to another, as "agenesic." Nevertheless, the writer would not necessarily deny the value of those formulations which look on these cases as cases of faulty adaptation to environment.16. As to (b), the functional correlations of this study, the resuits may be summed up by saying that strong correlations have been found to support the writer's former claims that (1) delusions are as a rule based on frontal disease, and (2) catatonic symptoms on parietal-lobe disease. An equally strong correlation (3) has now been found between auditory hallucinosis and temporal-lobe disease.17. The writer's previous work had suggested a correlation beteween frontal-lobe disease and delusion-formation. This correlation is not so decided in the present series, since, although perhaps only one of the 25 cases failed to exhibit delusions, seven of the remaining 24 failed to show frontal-lobe lesions. However, two of these seven, though grossly negative, were microscopically positive enough.18. The findings indicate, accordingly, that there is a group of delusional cases such that even long duration does not determine a frontal emphasis of lesions. Five cases represent this exceptional condition: three of these five are probably best interpreted as cases of hyperphantasia in which, both a priori and by observation, frontal lesions are not characteristic.19. On the whole, the correlation between delusions and focal brain atrophy (or aplasia capped by atrophy?) is very strong, particularly if we distinguish (1) the more frequent form of delusions with frontal-lobe correlations from (2) a less frequent form with parietal-lobe correlations.20. The non-frontal group of delusion-formations, the writer wishes to group provisionally under the term hyperphantasia, emphasizing the overimagination or perverted imagination of these cases, the frequent lack of any appropriate conduct-disorder in the patients harboring such delusions, and the a priori likelihood that these cases should turn out to have posterior-association-center disease rather than disease of the anterior association-center. This anatomical correlation is in fact the one observed.21. The writer's previous work had suggested a possible correlation between catatonic phenomena and parietal (including postcentral) disease: 10 of 14 definitely catatonic cases yielded parietal or other post-Rolandic lesions; two were grossly negative but microscopically altered; and indications of correlation appeared also in the remaining two. Five of seven clinically some what doubtfully catatonic cases yielded similar correlations. Four clinically non-catatonic cases yielded no parietal correlations. (It is worth while insisting that "catatonia" is here used to refer to a symptom, not to an entity or clinical group.)22. Special interest attaches to cerea flexibiita as a clearly definable form of catatonic symptom: four of five cases yielded gross parietal lesions. The fifth case was one of the entirely negative cases in the gross, but showed very marked postcentral satellitosis microscopically. Two of these cases showed the gross emphasis of lesions in the postcentral gyri, thereby hinting at an explanation of cerea flexibilitas along the lines of a reaction to altered kinæsthesia or an altered reaction to normal kinæsthesia (depending upon such true analysis of intragyral cortex-function as the future may bestow).23. À priori one might expect a correlation between the characteristic auditory hallucinosis found in many cases of dementia præcox and temporal-lobe lesions. In point of fact, nine of 12 hallucinated cases yielded temporal-lobe atrophy or aplasia; and actually only one of the three others is a good exception to the rule (from the clinical standpoint), to say nothing of the fact that this case had ample microscopic changes in the temporal lobe.24. Of the 13 non-hallucinating (auditory) cases, only three, or at most four, could be said to have temporal-lobe lesions suggesting the possibility of hallucinosis; here we may appeal to the inadequacy of clinical work, or, better, to the non-suitability of the lesions, since no one would assert that we yet have any idea of the precise and intimate temporal-lobe conditions which permit hallucinations.25. In these functional connections, the more recent formulations of Kraepelin and of Bleuler have been reviewed, although the entire work was done without the benefit of their analyses. The present formulation appears consistent enough with either. It would seem that Kraepelin regards a correlation between auditory hallucinosis and temporal-lobe disease as already highly probable from the literature. He also goes so far as to incriminate the "central" region for motor disorders. But the present suggestions as to the possible kinlesthetic relations of catatonia and the special (frontal and parietal) correlations with delusionformation are not suggested by Kraepelin from the literature available.26. It is interesting to note that further study by the Munich workers seems to have drawn attention away from the infrastellate cortical changes sketched by Alzheimer for catatonia in 1897 to various suprastellate changes. The microscopic work done in the present study in connection with certain grossly negative cases indicates that the early phases of the process may very often look as if infrastellate change was to be the most striking product of the disease. This is perhaps due to a richer original supply of glia cells in these infrastellate layers. Later, when the process is less acute, it may often be found that suprastellate cell losses are much more in evidence than any striking infrastellate change.27. As for the general position which this work would assume toward the functional conclusions of Bleuler, it would seem that a histopathological basis for "dissociations" or "schizophrenia" could be somewhat readily provided by the lesions found, since these are for long periods mild enough and sufficiently confined to the finer cortical apparatus to provide for the exquisite mental changes of most cases. The main neuronic systems are often permanently preserved, leaving an irregularly and slightly simplified cortical apparatus, in which a few cell changes would naturally throw out of coordination a great deal of still intact apparatus. But the whole process often remains so mild as to permit reestablishment of relatively normal functional relations on a slightly simplified basis, the whole to be disturbed once more on the occasion of the death or disease of a few more cells. Very striking is the fact that the cells not attacked are, so far as we can see, normal enough.28. This work is rather a study of genesis than of etiology, in the sense of modern medical distinctions between these branches of inquiry. It is a modest inquiry into factors, and does not rise to the height of ascribing causes. The writer will refer merely to some paragraphs in the text as to a possible ontological position concerning structure and function which the future may take. The deplorable thing is that some structuralists throw out of court all functional data and some (rather more!) functionalists tend to underrate the possible contributions of anatomy to this field. Luckily, science nowadays cannot long proceed merely á la mode.29. In particular, to sum up, I would call especial attention to the following points: (1) The constancy of mild general or focal atrophies in cases lasting long enough to yield these; (2) the tendency to an exhibition of lesions somewhat more markedly in the left hemisphere; (3) the preference of the lesions for the " association-centers" of Flechsig; (4) the high correlation of auditory hallucinosis and temporal-lobe lesions, as also (5) of catatonia and parietal lesions (cerea flexibilitas, especially postcentral), and (6) of the more frequent form of delusions and frontal-lobe disease; (7) the possible existence of a hyperphantasia group with parietal correlations, and of (8) a large internal hydrocephalus group with catatonic and hallucinotic correlations rather than delusional. A few more points can be got from the description of the accompanying plates.