THE CEREBROSPINAL FLUID 
 
 BY 
 
 LEWIS H. WEED 
 
 The Anatomical Laboratory, Johns Hopkins University 
 
 REPRINTED FROM PHYSIOLOGICAL REVIEWS 
 Vol. II, No. 2, April, 1922 
 
4-. 
 
Reprinted from PHYSIOLOGICAL, REVIEWS 
 Vol. II, No. 2, April, 1922 
 
 THE CEREBROSPINAL FLUID 
 
 LEWIS H. WEED 
 The Anatomical Laboratory, Johns Hopkins University 
 
 It is obviously impossible within the limitations of this review to 
 present a truly comprehensive account of a characteristic body-fluid such 
 as the cerebrospinal liquid. During the last century many contribu- 
 tions to our knowledge of this fluid have been made, though rather 
 sporadically and with long intervals between publications. Anatomists, 
 physiologists, pathologists and other workers have studied the problems 
 of this fluid; the proper presentation of the many phases of the subject 
 would lead into all of the representative biological sciences. But during 
 the past two decades contributions of a physiological and anatomical 
 nature have resulted in definite enlargement of our conceptions of this 
 fluid which so completely fills the cerebral ventricles and surrounds the 
 central nervous system. It is largely with these more recent advances 
 in knowledge that this review will deal, taking from the older literature 
 only those contributions which have founded the essential bases of the 
 biological processes of the cerebrospinal fluid. It is purposed to omit 
 in large measure the exact chemical, pathological and serological aspects 
 of the subject except as the data from these investigations aid in estab- 
 lishing the fundamental anatomical and physiological phenomena which 
 have to do with this fluid. For in this problem, as in many others, it 
 has seemed obvious that the furtherance of investigations upon function 
 has depended largely upon equivalent advance in anatomical knowledge. 
 
 The cerebrospinal fluid, as first effectively described by Magendie 
 (48), is a clear limpid liquid of low specific gravity (1.004 to 1.006), 
 colorless, and of a slight but definite viscosity. When withdrawn during 
 life, the liquid usually contains but few cells per cubic millimeter (less 
 than 10) but in many pathological conditions its cell-content may be 
 enormously increased. Various estimates of the amount of the fluid 
 existing in the cerebral ventricles and about the nervous system in 
 adult man have been published; the computation of 100 to 150 grams 
 
 171 
 
 PHYSIOLOGICAL REVIEWS, VOL. II, NO. 2 
 
 737135 
 
172 LEWIS H. WEED 
 
 given by Testut (65) is probably the most reliable, but because of the 
 complexities of the fluid-bed the figiires should necessarily be taken as 
 an approximation. 
 
 Chemifea-1 'examinations have demonstrated that the cerebrospinal 
 ;'fiiiid ;<?(intltiris small quantities of inorganic salts, of protein and of 
 ' dextrose; ' ' The Inorganic salts are chiefly sodium chloride and potas- 
 sium chloride; these occur in the ratio of 17.3 to 1, according to Mestre- 
 zat (50), whose monograph is a compendium of analyses of both normal 
 and pathological fluids. The average pH value is given by Felton, 
 Hussey and Bayne-Jones (27) as 7.75, with maximum variations of 7.4 
 and 7.9. Such chemical and physical characters induced Halliburton 
 (37) to term the cerebrospinal fluid "an ideal physiological saline solu- 
 tion," bathing the neurones and maintaining their osmotic equilibrium. 
 This liquid, then, of distinctive constitution,- unlike other body-fluids 
 (except the aqueous humor of the eye) becomes the subject of review. 
 Definite conceptions regarding its circulation are current, for the evi- 
 dence today points to a constant production of the fluid, its passage 
 through the cerebral ventricles and thence throughout the subarachnoid 
 space, and its subsequent major absorption into the venous system. 
 Necessarily, however, these conceptions must be subjected to critical 
 analysis, to determine the character and reliability of the data support- 
 ing these viewpoints. For this purpose it becomes essential to arrive 
 at some understanding of the anatomical mechanisms involved and to 
 ascertain the functional employment of these structures as pathways for 
 the cerebrospinal fluid. Such a discussion, while in no way comprehen- 
 sively covering the problems of the cerebrospinal fluid, will present at 
 least in part the phases of the subject which have in recent years been 
 most extensively investigated. 
 
 THE SOURCES OF THE CEREBROSPINAL FLUID 
 
 The description of the glandular histological structure of the choroid 
 plexuses by Faivre (25) in 1853 marked the discarding of the older con- 
 cept of Haller and Magendie that the cerebrospinal fluid was a product 
 of the leptomeninges (particularly of the pia mater) . Faivre made the 
 first histological survey of these villous projections, showing that the 
 cell coverings were epithelial in nature and that there were indications of 
 secretory activity in these cells. Faivre's observations, supported by 
 similar histological descriptions of hyaline-like inclusions by Luschka 
 (47) in 1855, gave origin to the hypothesis that the choroid plexuses 
 elaborate the greater portion of the cerebrospinal fluid; this has remained 
 
THE CEREBROSPINAL FLUID 173 
 
 the hypothesis upon which most of the investigations regarding the 
 source of this peculiar body-fluid have been based. 
 
 The purely histological evidence presented in support of this hypoth- 
 esis, while suggestive, lacks the element of conclusive proof, though 
 for many years accepted without question. During the past twenty 
 years, however, renewed attempts with finer histological and cytological 
 methods have been made to bridge the gap between intracellular secre- 
 tion-granules (vacuoles) and the actual production of the liquid sur- 
 rounding the cells. Thus, Findlay's (28) description of the granular 
 structure of the normal epithelial cells of the plexus, with frequent in- 
 clusions of slightly pigmented globules representing the fusion of smaller 
 elements and staining with osmic acid, is quite typical of the purely 
 histological demonstration of secretion. Studnicka (64) obtained some- 
 what similar evidence of secretion by the cells of the plexus and by the 
 ependymal cells of certain areas. Pettit and Girard (55) found hyaline- 
 like globules, similar to those described by Luschka, in the cells of the 
 plexuses in a comprehensive series of animals but did not feel that they 
 represented secretion-vacuoles. Loeper (46), working on the plexuses 
 in man, described pigmented granules and other granules within vacuoles 
 staining with osmic acid ; he stated that he believed that such histological 
 findings permitted him to assert that the cells of the choroid plexuses 
 are glandular. Employing methods of supravital staining in addition 
 to the ordinary histological procedures, Schlapfer (61) concluded that the 
 protoplasm of the cells of the choroid plexus contained " globoplasten" 
 surrounded by a lipoid-like capsule; his histological findings offer but 
 little additional support for his assertion that the choroid plexuses se- 
 crete the cerebrospinal fluid. Galeotti's (35) description in rabbit and 
 mouse of three different intracellular inclusions (hyaline droplets, fuch- 
 sinophilic granules and small basophilic plasmosomes) afforded evidence 
 of cell-activity by the choroidal epithelium but did not demonstrate the 
 elaboration of the liquid by these structures. The same statement may 
 be made of Francini's (32) differentiation of two forms of secretion- 
 phenomena in these cells droplets formed in the cytoplasm and gran- 
 ules derived from the cell-nucleus. Engel's (23) demonstration of a 
 fuchsinophilic granule and a basophilic granule staining with methyl 
 green is purely histological evidence of intracellular inclusions but the 
 great variability in position of the granules can hardly be interpreted 
 as indicating different phases in the secretion-process. And there is 
 likewise no conclusive proof of function in Hworostuchin's (41) descrip- 
 tion of the changes in form of the mitochondria of the choroidal epithe- 
 
174 LEWIS H. WEED 
 
 Hum as showing that the cells play an active role in the elaboration of 
 cerebrospinal fluid. Yoshimura (80), in a histochemical investigation 
 of the cells of the plexus, believed that the complicated process of secre- 
 tion of the cerebrospinal fluid was related to the aggregation of the 
 smaller cytoplasmic granules together into larger vacuoles for discharge. 
 Pellizzi's (54) hypothesis that the epithelial cells of the plexuses secreted 
 granules which increased in size by absorption of the fluid-plasm until 
 extruded, was based on a comprehensive study of the plexuses in verte- 
 brates and may likewise be considered as suggestive support of the 
 general thesis. 
 
 It seems clear from the observations just detailed that while the many 
 workers upon the histological structure of the choroid plexuses have 
 described certain granules and vacuoles in these epithelial cells, there is 
 no conclusive evidence that these intracellular structures constitute the 
 intracellular mechanism for the elaboration of cerebrospinal fluid. The 
 difficulty of final demonstration that these granules are discharged into 
 the fluid or are in some way dissolved in that fluid, cannot at present be 
 surmounted. 
 
 But fortunately observations of a far more conclusive nature have 
 been made by a combination of pharmacological and histological 
 methods. Cappelletti (9) reported in 1900 that ether and pilocarpine 
 increased the flow of cerebrospinal fluid from an experimental fistula 
 and that atropine and hyoscyamine diminished it. While considering 
 that the differences in vascular reaction to these drugs might account 
 for the phenomena observed, Cappelletti felt that the action of pilo- 
 carpine as a general stimulant of gland-activity justified the assumption 
 that there was a true acceleration of secretion of the cerebrospinal fluid. 
 Pettit and Girard (55) extended these observations of Cappelletti by 
 introducing histological examinations of the choroid plexuses in animals 
 which had been given muscarin, pilocarpine, ether, theobromine, etc. 
 The administration of these substances was found to increase the volume 
 of the cytoplasm of the choroidal epithelium, so that the cells became 
 doubled in height. Histological study of these enlarged elements, both 
 in the fresh and in fixed material, showed that the cells were divided into 
 a densely granular basilar zone and a clear apical zone. While the clear 
 apical area was indicated in the resting cell, the rapid enlargement of 
 this zone, under the influence of muscarin, ether and theobromine, 
 resulted in the formation of a distal clear vesicular mass. Such a dem- 
 onstration of histological change in the choroidal epithelium under 
 the influence of drugs which caused an increased flow of cerebrospinal 
 
THE CEREBROSPINAL FLUID 175 
 
 fluid from a fistula, was believed by Pettit and Girard to prove conclu- 
 sively that the choroid plexuses possessed a secretory function in the 
 elaboration of cerebrospinal fluid. 
 
 Meek (49), repeating the experiments of Pettit and Girard, came to 
 identical conclusions. Muscarin caused no change in the choroidal 
 epithelium in the rat but in the rabbit and guinea pig definite alterations 
 were recorded. Meek stated (p. 300) that "the two things most 
 striking about these modified cells are their great increase in height and 
 the appearance of so much clear space at the distal side of the nucleus." 
 
 While the observations just quoted definitely relate the choroid 
 plexuses to the elaboration of cerebrospinal fluid, there is available other 
 substantiating evidence in support of this hypothesis. It had long been 
 realized, from pathological examinations of cases of obstructive internal 
 hydrocephalus, that the cerebrospinal fluid must be at least in part 
 produced by some intraventricular structure. In this relationship 
 renewed attention was directed to the choroid plexuses by the discovery 
 by Claisse and Levy (12) in 1897 of a case of internal hydrocephalus asso- 
 ciated with hypertrophy of these intraventricular plexuses. Dandy and 
 Blackfan (17), (18) and Frazier and Peet (33) gave additional support 
 to the general contention when they experimentally produced an internal 
 hydrocephalus by occlusion of the aqueduct of Sylvius. Cushing's (14) 
 observation of an exudation of a clear fluid from a choroid plexus ex- 
 posed in exploration of a porencephalic defect likewise added suggestive 
 substantiation of the hypothesis. Somewhat more tangible proof of 
 intraventricular elaboration of the fluid was afforded by the writer's (68) 
 demonstration that a definite and sustained outflow of cerebrospinal 
 fluid could be obtained by catheterization of the third ventricle through 
 the aqueduct of Sylvius. The outflow from such a catheter was quite 
 similar in amount to the fluid obtained from a cannula in the subarach- 
 noid space; the finding argues strongly for the belief that the major 
 portion of the cerebrospinal fluid is produced within the cerebral ven- 
 tricles. But Dandy's (16) later experiments constitute dependable 
 evidence not only that the place of production of cerebrospinal fluid is 
 intraventricular but that the choroid plexuses are the responsible struc- 
 tures. Dandy was able to produce a unilateral internal hydrocephalus 
 by obstructing one foramen of Monro; extirpation of the choroid plexus 
 in such an obstructed lateral cerebral ventricle prevented the develop- 
 ment of an internal hydrocephalus. Dandy's experiment furnishes 
 the strongest single substantiation of the hypothesis that the choroid 
 plexuses elaborate the cerebrospinal fluid. 
 
176 LEWIS H. WEED 
 
 From an entirely different aspect, also, corroborative evidence in 
 favor of the choroidal origin of the cerebrospinal fluid is found in em- 
 bryological observations of the writer (69). In a study of the develop- 
 ment of the cerebrospinal spaces it was shown that the first extraven- 
 tricular expulsion of the cerebrospinal fluid occurred simultaneously 
 with the first tufting and histological differentiation of the ependymal 
 cells to form the choroid plexuses. 
 
 The evidence, then, from histological, pharmacological, pathological 
 and embryological standpoints, surely inclines one to acceptance of the 
 hypothesis that the choroid plexuses of the cerebral ventricles largely 
 elaborate the cerebrospinal fluid. It does not seem justifiable to accept 
 the evidence from any one standpoint as conclusive for many of the 
 observations recorded are corroborative only. Yet the general histo- 
 logical structure of these plexuses, the cytoplasmic inclusions, and the 
 modification of the cell-structure by pharmacological agents offer more 
 than suggestive substantiation of the contention. The pathological 
 studies of cases of internal hydrocephalus, the direct observations of the 
 "sweating" choroid plexus, the embryological relationship between 
 differentiation of choroid plexuses and extraventricular spread of the 
 ventricular fluid, and particularly the experimental investigation of 
 unilateral hydrocephalus, when considered as a whole, present a very 
 strong argument, if not wholly conclusive, in favor of the view that the 
 choroid plexuses are the elaborators of the major portion of the cere- 
 brospinal fluid. It does not seem justifiable to discard, as Becht (1) 
 has done, all of the evidence in favor of this hypothesis as inconclusive. 
 While many of the experimental findings, when viewed as isolated 
 observations, may be explained by other hypotheses, the great mass of 
 data cannot be interpreted on any other hypothesis as satisfactorily. 
 Certain of Becht's specific objections to acceptance of the theory of 
 origin of the liquid from the choroid plexuses have been answered within 
 the last year: Wislocki and Putnam (79) demonstrated by histological 
 methods that in cases of experimental internal hydrocephalus there was 
 absorption of foreign solutions through the ependymal cells lining the 
 ventricles but not through the cells of the choroid plexus. These ob- 
 servations were confirmed and extended by Nanagas (53), who showed 
 by similar procedures that an increased intraventricular absorption of 
 fluid occurred after the intravenous injection of hypertonic solutions of 
 sodium chloride ; in no case was there absorption of the fluid by the cells 
 of the choroid plexus. With this evidence in hand, it seems justifiable 
 to disregard Becht's contention that the cell-changes in the choroid 
 
THE CEREBROSPINAL FLUID 177 
 
 plexuses, reported by Pettit and Girard and by Meek, may as properly 
 be interpreted as indicative of absorption as of secretory activity. 
 
 But even as a working hypothesis, the choroid plexuses must not be 
 considered to be the sole elaborators of the cerebrospinal fluid. Ana- 
 tomical evidence presented by the writer (68) indicates that the perivas- 
 cular spaces also pour a certain amount of fluid into the subarachnoid 
 space, where this fluid mixes with the liquid produced in the cerebral 
 ventricles. Such an addition to the cerebrospinal fluid probably ac- 
 counts for the reported differences between subarachnoid and ventric- 
 ular fluids on serological and chemical analysis. The ependymal cells 
 lining the cerebral ventricles and the central canal of the spinal cord 
 may also contribute even in the adult a minimal addition to the intra- 
 ventricular cerebrospinal fluid. 
 
 Although a constant elaboration of cerebrospinal fluid by these 
 mechanisms is indicated, it is not known how large a quantity is pro- 
 duced in any given time-interval but it is not unlikely that the majority 
 of computations are by far too large. Most of the estimates in man 
 have been based on the amount of fluid pouring from subarachnoid 
 fistulae (also cases of cerebrospinal rhinorrhea) where the pressure, 
 against which the fluid is produced, is determined solely by the resist- 
 ance of the abnormal pathway. The same lack of normal conditions 
 renders unreliable the determinations which are based on the rate of 
 flow from experimental cannulae or fistulae. Estimations of the produc- 
 tion of fluid based on the absorption of foreign dyes likewise may lead 
 astray. The evidence indicates, however, that there is a constant 
 though not excessive elaboration of the cerebrospinal fluid; the com- 
 putations of exact quantities thus far given are of but little value. 
 
 CIRCULATION OF THE CEREBROSPINAL FLUID 
 
 The cerebrospinal fluid produced largely by the choroid plexuses is 
 poured directly into the cerebral ventricles which are lined by ectodermal 
 ependymal cells. That portion of the fluid formed in the lateral ven- 
 tricles flows through the foramina of Monro into the third ventricle and 
 thence by the aqueduct of Sylvius into the fourth ventricle. From the 
 fourth ventricle the fluid passes out into the subarachnoid space; there 
 is no evidence that functional communications between cerebral ven- 
 tricles and subarachnoid space exist elsewhere than in this region. 
 
 The exact mode of escape of the ventricular cerebrospinal fluid from 
 the fourth ventricle into the subarachnoid space must still be considered 
 as slightly uncertain. It is possible that the inferior velum of the cere- 
 
178 LEWIS H. WEED 
 
 bellar roof in the adult is an intact though functioning membrane, as in 
 the embryo; the existence of a perforation (the foramen of Magendie) 
 in this membrane has been termed an artifact because of the dislocation 
 of structures necessary to demonstrate it macroscopically or because of 
 the shrinkage of tissues in embedding for histological investigation. The 
 greater weight of evidence today inclines to a consideration of the fora- 
 men of Magendie as a true anatomical opening in the velum a break 
 in the ependyma and pia. In support of this conception of a true fora- 
 men between fourth ventricle and subarachnoid space may be quoted 
 the observations of Cannieu (8) and of Wilder (78) and especially the 
 developmental studies of Hess (38) and of Blake (6). Blake's concep- 
 tion of the formation of this opening a shearing-off of the base of a 
 finger-like evagination of the rhombic roof is rendered more certain 
 by recent morphological studies of this region. The two lateral 
 foramina those of Luschka connecting the lateral recesses of 
 the fourth ventricle with the subarachnoid space, seem to have as 
 established a basis for their existence as does the medial. It is 
 through these three foramina or surely in the region of the inferior 
 tela choroidea if through an intact membrane that the cerebrospinal 
 fluid, produced in the cerebral ventricles, passes into the subarachnoid 
 space. 
 
 From the cisternal dilatation of the subarachnoid space in the region 
 of the medial cerebello-bulbar angle the cerebrospinal fluid slowly seeps 
 downward in the spinal subarachnoid space but passes more rapidly 
 upward about the base of the brain and thence more slowly over the 
 hemispheres, surrounding the whole central nervous system. This 
 movement of fluid is facilitated by impulses transmitted to it by the 
 vascular system; in the spinal region there is also an almost equivalent 
 passage of fluid upward. The subarachnoid space, in which the fluid 
 circulates, is, according to current anatomical descriptions, contained 
 between the arachnoidea and the pia mater. Such a description does 
 not present a proper conception of these fluid-containing channels, for 
 it seems far preferable to consider the subarachnoid space to be intra- 
 leptomeningeal. On this basis the arachnoid may be described as the 
 outer continuous membrane, intact and fluid-containing, from the inner 
 surface of which project numerous delicate trabeculae, which merge 
 with the pia mater. The surfaces of the arachnoid membrane and of the 
 trabeculae are covered, as is the inner surface of the dura mater, by 
 flattened, polygonal mesothelial cells. Identical cells also clothe the 
 surface of the brain and spinal cord to form the essential cell-covering 
 
THE CEREBROSPINAL FLUID 179 
 
 of the pia mater. All structures (blood vessels, nerves, etc.) traversing 
 the subarachnoid space are likewise covered by these mesothelial ele- 
 ments. The cerebrospinal fluid, hence, is contained within a completely 
 cell-lined system of continuous yet partially interrupted spaces in the 
 meshes of the arachnoid trabeculae. These meshes are of various sizes, 
 increasing from very fine reticular spaces over the cerebral hemispheres 
 to more widely calibered channels in the cerebral sulci and about the 
 spinal cord, and reaching their maxima in the cisternal dilatations about 
 the cerebello-bulbar angle. In the wide channels of this subarachnoid 
 meshwork the cerebrospinal fluid is obstructed but little in its circula- 
 tion, but in the small meshes the flow of the liquid is slowed. 
 
 Apart from their established function as efficient fluid-retainers, the 
 cells lining the subarachnoid space are of great interest because of their 
 changing morphology under different physiological conditions. The 
 writer (70) first noted that these mesothelial cells phagocyted carbon 
 granules introduced into the subarachnoid space and that when phagocy- 
 tic, the cells increased in size. Essick (24) then showed that the pres- 
 ence of foreign material (laked blood, granules, etc.) caused these cells 
 to become enlarged, vacuolated, phagocytic and finally detached to 
 form free macrophages of the cerebrospinal fluid. 
 
 These mesothelial cells likewise have importance in establishing the 
 relations between the subarachnoid and the perivascular spaces, for there 
 is everywhere in the central nervous system a distinct fluid-containing 
 space about each of the perforating blood vessels. The cells of the pia 
 mater turn inward to form an outer wall of such a perivascular channel 
 while the cells of the arachnoid, covering the vessel as it traverses the 
 subarachnoid space, are likewise continued inward to form an inner cuff 
 of this space. Thus each blood vessel penetrating the nervous system 
 is surrounded by a cell-enclosed, peri-adventitial fluid-channel, which 
 communicates directly with the subarachnoid space. The typical 
 leptomeningeal mesothelial cell of this channel has been identified for 
 variable distances from the surface, dependent upon the caliber of the 
 penetrating vessel. The perivascular fluid-channel, when the mesothe- 
 lial cell-cuff ceases, continues inward to connect directly with perineuronal 
 spaces about the nerve-cells. These ultimate fluid-spaces are potential 
 in character but under certain circumstances they become easily rec- 
 ognizable in microscopic preparations (Mott (52)). While originally 
 termed "lymphatic" in character, these perivascular channels have no 
 connection with the lymphatic system; they represent however an im- 
 portant accessory fluid-system of the cerebrospinal axis, affording direct 
 
180 LEWIS H. WEED 
 
 pathway, uninterrupted by cell-membranes, between nerve-cell and 
 subarachnoid space. 
 
 Thus far mention of the dura mater has been omitted, for it has but 
 slight relationship to the cerebrospinal fluid, the subdural space being 
 anatomically and probably physiologically entirely apart from the sub- 
 arachnoid. But in one respect the dura mater has importance in the 
 present problem : the areas of penetration of the dense fibrous tissues of 
 the dura by the arachnoid represent points of fusion between the two 
 membranes. The most frequent of these areas of penetration are the 
 arachnoid villi prolongations of the arachnoid membrane so that the 
 arachnoidal mesothelial cells come to be directly beneath the vascular 
 endothelium of the great dural venous sinuses. Identified in adult man, 
 in infants and in the common laboratory mammals (Weed (66), (67)), 
 these villi are covered by typical arachnoid cells, usually of a single 
 layer but often forming whorls and presenting double-layered coverings. 
 The core of such a villus may be a strand-like network reduplicating the 
 subarachnoid space or a myxomatous ground work simulating the 
 perimedullary niesenchyme. In addition to the true arachnoid villi, 
 which occur in the walls of practically all of the dural sinuses, there are 
 found infrequently prolongations of the arachnoid into the dura in other 
 situations. The arachnoid villi are normal structures; the great enlarge- 
 ment of these in adult life results in the formation of the well-known 
 Pacchionian granulations. 
 
 The cerebrospinal fluid, then, circulates everywhere about the central 
 nervous system in the cerebral ventricles and central canal of the 
 spinal cord and also in the tortuous meshes of the subarachnoid space. 
 These channels are all clothed with a specialized cell, fluid-retaining so 
 that a true circulation of fluid may be maintained. And in the arach- 
 noid villi the circulating fluid comes into close relationship to the great 
 venous sinuses of the dura mater. 
 
 ABSORPTION OF THE CEREBROSPINAL FLUID 
 
 With the evidence indicating a constant production of cerebrospinal 
 fluid and a circulation of the liquid through the cerebral ventricles and 
 throughout the subarachnoid space, it is not surprising that many investi- 
 gations should have been undertaken to determine the mode of absorp- 
 tion of the fluid. The experiments performed fall naturally into two 
 groups the physiological observations to determine whether absorption 
 of the fluid is into venous system or lymphatic trunks and the anatomical 
 investigations to ascertain the exact pathways along which the fluid is 
 
THE CEBEBROSPINAL FLUID 181 
 
 absorbed. Because of the great importance of this phase of the subject 
 the evidence will be given in some detail. 
 
 Modern anatomical studies of the pathways of absorption of the cere- 
 brospinal fluid were first made by Key and Retzius (43) who injected 
 into the spinal subarachnoid space of a cadaver a gelatine solution 
 colored with Berlin blue. While the pressure used was somewhat exces- 
 sive (60 mm. Hg.), the anatomical continuity of spinal and cranial sub- 
 arachnoid spaces was demonstrated, for the whole cranial subarachnoid 
 space was filled with the gelatine-mass. Furthermore, the gelatine was 
 traced into the core of the Pacchionian granulations along the great 
 dural sinuses, and through the cell-membranes directly into these 
 venous sinuses. Many beautiful plates showing this passage of the 
 injection-mass but giving evidence of possible rupture are presented in 
 Key and Retzius' monograph. In addition to this major pathway of 
 absorption, these investigators demonstrated a minor accessory absorp- 
 tion of the fluid into the lymphatic system. Quincke's (56) confirma- 
 tory observations, though appearing before the monograph of Key and 
 Retzius, did not antedate their earlier papers on the subject. Quincke 
 injected into the subarachnoid space of living animals a suspension of 
 cinnabar granules and, killing the animals at varying periods thereafter, 
 discovered the granules almost wholly within the basilar and spinal 
 subarachnoid spaces, for the most part held by phagocytic cells. 
 Granules were also found along the venous sinuses in structures which 
 he termed Pacchionian granulations; in the cervical lymph nodes like- 
 wise particles of the sulphide were identified. 
 
 For several years after these publications, the view of Key and Retzius 
 was accepted as establishing the anatomical pathway of absorption of 
 the cerebrospinal fluid. But gradually with the realization that Pac- 
 chionian granulations as such do not exist in infants and in the higher 
 mammals, it was felt that this view was inadequate. For the next two 
 decades practically no work was done upon the subject; then suddenly 
 renewed interest in the problem was made manifest by the publication 
 of several important physiological observations demonstrating a major 
 absorption of the liquid into the venous system and a minor lymphatic 
 drainage. 
 
 Reiner and Schnitzler (57) injected saline solutions containing potas- 
 sium ferrocyanide into the spinal subarachnoid space of living animals 
 and recovered the foreign salt in from 30 to 40 seconds from the blood 
 of the jugular vein. Olive oil injected under similar conditions was 
 identified likewise, though the blood stream was slowed. Reiner and 
 
182 LEWIS H. WEED 
 
 Schnitzler stated that as Pacchionian granulations do not exist in the 
 animals used, other pathways of absorption must exist. Shortly there- 
 after Leonard Hill (39) reported that saline solution, colored with 
 methylene blue and introduced into the subarachnoid space, could be 
 traced "straight into the venous sinuses." Within a few minutes (10 
 to 20) after the injection, the blue was identified in the bladder and 
 stomach; the cervical lymphatics became colored only after an interval 
 of one hour. Following injection of potassium ferrocyanide into the 
 cerebrospinal fluid, Ziegler (81) detected the foreign salt in the posterior 
 facial vein in 10 seconds and only after 30 minutes in the cervical 
 lymphatics. Similarly, Lewandowsky (44) identified sodium ferrocya- 
 nide in the urine of animals within 30 minutes after intraspinal 
 introduction. 
 
 The experiments of Spina (63), though conducted under abnormally 
 high pressures, likewise add support to the idea of a very rapid major 
 venous absorption and a lesser lymphatic drainage. Cushing's (13) 
 observations on mercury and non-absorbable gases led him to hypothe- 
 size a valve-like mechanism for drainage into the venous channel. Both 
 of these observers commented upon the absence of Pacchionian granula- 
 tions in the higher mammals. 
 
 Mott (52) in 1910, from study of the brains of animals subjected to 
 experimental cerebral anemia, suggested a new pathway for the absorp- 
 tion of the cerebrospinal fluid, based on the occurrence of distinct spaces 
 about each nerve-cell, connected through the perivascular channels 
 with the subarachnoid space. Believing that this fluid-system contained 
 cerebrospinal fluid, Mott contended that normally the liquid passes 
 from subarachnoid space into cerebral capillaries. Cathelin (10), with- 
 out supporting data, assumed that the major absorption of the cere- 
 brospinal fluid was into the lymphatic system; and Goldmann (36), 
 employing subarachnoid injections of trypan blue, likewise tentatively 
 favored this view, though acknowledging the weight of evidence in 
 favor of the major venous drainage. 
 
 Dandy and Blackfan (17) in 1913 concluded that the absorption of 
 cerebrospinal fluid was a "diffuse process from the entire subarachnoid 
 space," for with the spinal subarachnoid space isolated from the cranial, 
 they found "a quantitative absorption proportionately as great as from 
 the entire subarachnoid space." The evidence for these statements 
 (18), published in detail a year later, was largely based on the excretion 
 by the kidneys of a readily diffusible dye phenolsulphonephthalein 
 after its introduction into the subarachnoid space. A very rapid 
 
THE CEREBROSPINAL FLUID 183. 
 
 absorption into the blood stream occurred under such experimental con- 
 ditions while the lymphatic drainage was minimal in amount and very- 
 tardy. After subarachnoid injections of india ink, Dandy and Blackfan 
 were able to find no anatomical evidence of absorption of the carbon 
 granules an observation in accord with those of Quincke (56) and of 
 Sicard and Cestan (62). 
 
 It was with these contributions as a background that the writer (66), 
 (67), (68) began his anatomical studies of the absorption of the cere- 
 brospinal fluid. Critical examination of the preceding work was con- 
 vincing in demonstrating a major venous absorption and a minor lym- 
 phatic drainage of the liquid, but the reliable data were practically 
 entirely physiological. Thus the observations regarding the rapid 
 venous absorption of readily diffusible substances seemed wholly de- 
 pendable but the anatomical evidence was satisfactory solely in demon- 
 strating that insoluble particles (cinnabar, carbon) did not leave the 
 subarachnoid space in any great amount. The only complete morpho- 
 logical investigations were those of Key and Retzius; these were unsatis- 
 factory because the injections were performed on dead material, the 
 pressures employed were high (60 mm. Hg.) and the injection-mass was 
 a viscous colloid (gelatine) colored with Berlin blue. It was felt that 
 the experimental approach must be such that a subarachnoid injection 
 of a true, isotonic solution of non-toxic foreign salts, capable of subse- 
 quent precipitation in situ for histological examination and not diffusely 
 staining cellular material, could be made in the living animal, under 
 pressures not greatly in excess of the normal. With these criteria 
 established, experiments were undertaken with subarachnoid injection 
 of potassium ferrocyanide and iron-ammonium citrate in isotonic 
 solution under pressures but slightly above the normal (130-180 mm. 
 H 2 0). Subsequently the central nervous system, enclosed in meninges, 
 was fixed in an acid medium; precipitation of the foreign salts as Prus- 
 sian blue permitted adequate histological identification of the pathway 
 taken. This method was found to meet all of the standards of investi- 
 gation set. 
 
 The experiments were carried out over periods of several hours in 
 living anesthetized animals, with introduction of the isotonic foreign 
 solution into the lumbar subarachnoid space. The spinal and basilar 
 portions of the subarachnoid space were rapidly filled with the foreign 
 solution but the cerebral portion of the space was not completely in- 
 jected until the experiment had been continued for several hours. 
 Histological examination demonstrated that the solution had not pene- 
 
184 LEWIS H. WEED 
 
 trated any of the cells lining the subarachnoid space; the precipitated 
 granules adhered to the surfaces of the cells but were not within the 
 cytoplasm. The foreign solution was found to have passed directly 
 into the venous sinuses by way of the arachnoid villi into which the pre- 
 cipitated granules could be traced from the cerebral subarachnoid 
 space. These granules, representing the foreign solution, were found 
 within the mesothelial cells covering the villi and the endothelial cells 
 lining the venous sinuses, as well as within the lumen of the venous 
 sinus, thus demonstrating the essential pathway of absorption. In no 
 other place was there evidence of direct passage through a cell-mem- 
 brane as in the villi. The mechanism of passage of this fluid seemed 
 to be a process of filtration from a point of higher pressure (subarachnoid 
 space) to a point of lower pressure (venous sinus), though the factors of 
 osmosis and diffusion were not excluded. The absorption of true solu- 
 tions from the cranial subarachnoid space was shown to be a much 
 more efficient and rapid process than was the corresponding absorption 
 from the spinal subarachnoid space. Suspensions of particulate matter 
 were found, in agreement with the observations already recorded, to be 
 retained within the meshes of the subarachnoid space. 
 
 In addition to this major venous absorption through arachnoid villi 
 directly into the great dural sinuses, an accessory drainage by way of 
 the lymphatic system was demonstrated. This seemed a much slower, 
 less efficient means of absorption of the fluid, caring for but a small 
 fraction of the total. Such lymphatic absorption was wholly indirect; 
 the fluid reached the true lymphatic vessel only outside of the dura and 
 then by way of perineural spaces. 
 
 These anatomical findings, based on a standard of experimentation 
 which approximated the normal, agreed largely with those of Key and 
 Retzius, substituting however for the Pacchionian granulation the 
 normal arachnoid villus. Further observations were made to ascertain 
 the truth of the other anatomical hypotheses ventured for the absorption 
 of cerebrospinal fluid : thus, no structure of a valve-like nature was found 
 in many examinations of serial sections of the great dural sinuses. 
 Mott's theory of absorption by cerebral capillaries was excluded by the 
 failure of the injection-fluid to pass into the peri vascular system under 
 conditions approaching the physiological. Dandy and Blackfan's con- 
 ception of a diffuse absorption by the vessels of the subarachnoid space 
 was likewise found untenable, for in no case were the mesothelial cells 
 covering these vessels penetrated by the foreign solution. The strongest 
 argument in favor of the diffuse process of absorption advanced by 
 
THE CEREBROSPINAL FLUID 185 
 
 Dandy and Blackfan was that the excretion of phenolsulphonephthalein 
 from the isolated spinal subarachnoid space was proportionately as great 
 as from the whole cranial and spinal subarachnoid space. The tech- 
 nique employed by Dandy and Blackfan involved withdrawal of an 
 equivalent amount of cerebrospinal fluid from the isolated spinal sub- 
 arachnoid space before injection of 1 cc. of the phenolsulphonephthalein 
 solution. The writer (67) believed that this substitution could not be 
 made in this isolated space without increase in the spinal subarachnoid 
 pressure and escape of the foreign dye into the epidural tissues. By 
 reversing the experiment he was able to show that absorption of phenol- 
 sulphonephthalein when introduced into the cisterna magna was as 
 rapid when the spinal subarachnoid space was excluded as with the 
 whole cranial and spinal subarachnoid space functioning. The cranial 
 portion of the nervous system seems, therefore, to contain the efficient 
 mechanisms for the absorption of cerebrospinal fluid. 
 
 Since the publication of this work in 1914 there have been other 
 observations regarding the absorption of the cerebrospinal fluid; all 
 accord with this idea of a major venous absorption of the fluid. Thus 
 Frazier and Feet's (33) experiments with methylene blue, isamine blue, 
 trypan red, trypan blue and phenolsulphonephthalein demonstrated the 
 greater importance of the rapid venous absorption and the lesser of the 
 slow lymphatic drainage. And Dixon and Halliburton (21), in the 
 course of their studies of the cerebrospinal fluid, confirmed this general 
 conception of the process, finding no escape of particulate matter from 
 the subarachnoid space but a free and rapid absorption of true solutions. 
 Between the two types there occurred a much slower absorption of 
 colloidal solutions than of true solutions, the. larger molecules being 
 absorbed more slowly than the smaller. They concluded, as did also 
 Halliburton (37) that "the fluid probably reached the venous sinuses 
 by way of the microscopic arachnoid villi." And recently the writer, in 
 work as yet unpublished, has repeated certain phases of his original 
 investigation of the pathways of absorption, with continuous observa- 
 tions of the pressures of the cerebrospinal fluid and of the intracranial 
 blood vessels. The results confirm in every way the conception of the 
 mechanism here presented. 
 
 Thus it seems fair to assume that the absorption of the cerebrospinal 
 fluid is a twofold process, being chiefly a rapid drainage into the great 
 dural sinuses, and in small part a slow indirect escape into the true 
 lymphatic vessels. 
 
186 LEWIS H. WEED 
 
 THE PRESSURE OF THE CEREBROSPINAL FLUID 
 
 Practically all of the methods of recording the pressure of the cere- 
 brospinal fluid deal with connection of the subarachnoid space to an 
 open or membrane manometer. It is very difficult with any of the pro- 
 cedures to avoid the loss of a few drops of fluid, but by immediate re- 
 placement of this liquid and by the use of manometers filled to the esti- 
 mated level of the fluid, the pressures obtained by these simple methods 
 may be accepted as accurate. 
 
 Very few, if any, of the early records of the pressure of the cerebro- 
 spinal fluid were sufficiently controlled to permit direct comparison with 
 the data obtained by recent workers. As examples of the variation in 
 the pressures recorded by the earlier investigators, the following deter- 
 minations may be given: Key and Retzius (43), 162 to 216 mm. H 2 O 
 in inspiration and 216 to 270 mm. H 2 O in expiration, in etherized dogs; 
 Bergmann (4, 5), 80 mm. H 2 O in his first observations and in his second 
 series, 120 to 160 mm. of salt solution in narcotized dogs; Falkenheim 
 and Naunyn (26), 100 to 150 mm. H 2 O in curarized dogs; and L/eyden 
 (45), 80 to 150 mm. and 100 to 120 mm. H 2 in dogs under morphia. 
 Leonard Hill (39) believed, however, that the intracranial tension might 
 vary from zero to 50 mm. Hg. and that the variations reported by the 
 earlier observers were but expressions of this characteristic of the intra- 
 cranial pressure. 1 Hill's idea of variability in the normal pressure of 
 the cerebrospinal fluid seemed supported by the observations of Dixon 
 and Halliburton (20), who reported 40 to 70 mm. of salt solution as a 
 rough average of the normal pressures obtained in the dog under mor- 
 phine-urethane anesthesia. And in a single detailed protocol presented, 
 the pressures at 5-minute intervals ranged as follows : 95, 25, 30, 35, 55, 
 25, 80, 65, 65, 65, 75, 70, 60, 55, 50, 80, 90 mm. of 10 per cent citrate 
 solution. 
 
 The most recent work, particularly by American investigators, has 
 given much better knowledge of the range and variability of the normal 
 pressure of the cerebrospinal fluid. Weed and McKibben (75) reported 
 an initial average of 119 mm. of Ringer's solution in cats anesthetized 
 by intratracheal ether, and an extreme constancy of the pressure of the 
 cerebrospinal fluid under such experimental conditions. Becht (1) 
 found considerable fluctuation in the pressure of the fluid in etherized 
 dogs but of lesser extent than did Dixon and Halliburton; his average 
 
 1 Within the limits of the physiological phenomena investigated, intracranial 
 pressure may be considered to be identical with that of the cerebrospinal fluid. 
 
THE CEREBEOSPINAL FLUID 187 
 
 pressure of 112 mm. (sodium chloride solution of specific gravity of 
 1.088) was derived from 39 dogs under intratracheal ether. Foley and 
 Putnam (31) presented a slightly higher average of 127 mm. for the 
 normal reading in animals under ether while the average of 100 animals 
 under various anesthetics was 133 mm. of cerebrospinal fluid; no com- 
 ment upon the extent of normal fluctuations of the fluid pressure was 
 made. And recently Weed and Hughson (72) have reported an average 
 pressure of 119 mm. of Ringer's solution for 77 cats under ether by 
 Woulfe bottle; the fluctuations in the pressure under the experimental 
 conditions were very slight in extent (11 mm. in an animal under ob- 
 servation for 2 hours). 
 
 The question of greatest physiological interest in this phase of the 
 general problem is that of the maintenance of this pressure of the cere- 
 brospinal fluid. All of the conceptions of the mechanism for the main- 
 tenance of this pressure are based primarily upon the rigid character 
 of the bony coverings of the nervous system. This idea that the cere- 
 brospinal axis is situated within a "closed box," to which the physical 
 laws of such a system apply, was first advanced in 1783 by Alexander 
 Monro (51). Monro believed that the substance of the brain, like that 
 of other solids of the body, is nearly incompressible and is "enclosed 
 in a case of bone," assuring therefore the constancy of the intracranial 
 blood content. The development of this hypothesis by Kellie (42) in 
 1824 led to wide acceptance, and the Monro-Kellie doctrine with but few 
 alterations has served as the basis upon which the physiology of the 
 intracranial contents has been interpreted. The doctrine was modified 
 by the introduction of the cerebrospinal fluid into the consideration by 
 Burrows (7) under the influence of Magendie's epochal studies; the 
 hypothesis was then formulated by Burrows as follows (p. 32): "the 
 whole contents of the cranium, the brain, the blood and this serum 
 (cerebrospinal fluid) together, must be at all times nearly a constant 
 quantity." 
 
 Many physiologists have subjected this Monro-Kelllie doctrine to 
 experimental proof not only in dead but in living animals; while diver- 
 gent conclusions have been arrived at, the general consensus of opinion 
 expressed in the literature of the last century has been in favor of the 
 hypothesis. And in the hands of recent workers a similar unanimity 
 of expression holds though occasional investigators express doubt as 
 to the accuracy of the premise. Lately, Weed and Hughson (73) have 
 experimentally demonstrated the essential truth of the doctrine that 
 (p. 99) "the bony coverings of the central nervous system constitute 
 
188 LEWIS H. WEED 
 
 within tested physiological limits, inelastic and rigid containers; the 
 ordinary physical laws of a 'closed box' may therefore be applied to the 
 cranium." And with appreciation of the variability in volume of the 
 constituents, the hypothesis may be stated, as was done by Weed and 
 McKibben (76, p. 553) : "the cranial cavity is relatively fixed in volume 
 and is completely filled by brain, cerebrospinal fluid and blood; varia- 
 tions in any one of the three elements may occur, compensation being 
 afforded by alteration in the volume of one or both of the remaining 
 elements." 
 
 With the cranium and vertebral column serving as rigid containers, 
 the relation of the intracranial vascular pressures to the pressure of the 
 cerebrospinal fluid requires immediate consideration. Current physio- 
 logical conceptions of these intracranial mechanisms really date from 
 the work of Leonard Hill (39) who advanced the idea that intracranial 
 pressure is (p. 71) "the same as cerebral venous pressure, and varies 
 in the same direction absolutely as general venous pressure, and pro- 
 portionately as general arterial pressure." Hill's technical procedures 
 consisted in determing cerebral venous pressure in the torcular Hero- 
 phili and cerebrospinal fluid pressure in the occipital region. 
 
 The emphasis placed by Hill upon this equality of pressure dominated 
 physiological opinion for over fifteen years; it was not until the publica- 
 tion of Dixon and Halliburton (20) in 1914 that contradictory evidence 
 was presented. Using experimental procedures essentially similar to 
 those of Hill, Dixon and Halliburton demonstrated that the cere- 
 brospinal fluid pressure is not identical with that of the dural venous 
 sinus, and stated that the fluid pressure is (p. 153) "influenced passively 
 to a small extent by changes in the arterial and venous pressures but 
 such alterations are insignificant compared with the independent 
 changes in pressure which occur as the result of secretory activity." 
 These investigators also showed that increase in the cerebrospinal fluid 
 pressure produced a passive increase in the cerebral venous pressure 
 but not of the same extent; conversely, alteration of the cerebral venous 
 tension caused similar though not identical alteration in the pressure 
 of the cerebrospinal fluid. Under normal conditions the arterial pres- 
 sure was found by Dixon and Halliburton to be higher than the intra- 
 cranial venous pressure which in turn was always higher than that of the 
 cerebrospinal fluid. 
 
 Becht (1), employing somewhat similar methods, came to conclusions 
 in some respects at variance with those of Dixon and Halliburton though 
 confirming the general contention of inequality of cerebrospinal fluid 
 
THE CEREBROSPINAL FLUID 189 
 
 pressure and cerebral venous pressure. Becht stated that the cerebro- 
 spinal fluid pressure was dependent upon both intracranial arterial and 
 venous pressures, though not identical with either. The data obtained 
 indicated that either the cerebral venous (torcular) or cerebrospinal 
 fluid pressure might be the higher but that usually the former exceeded 
 the latter. Passive changes in the torcular pressure were found to affect 
 the pressure of the cerebrospinal fluid, in the same direction but not to 
 the same extent; but within fairly wide limits changes of the cerebro- 
 spinal fluid pressure were not believed to alter the pressure of the 
 torcular. 
 
 The observations of these recent workers were all carried out with 
 .fairly similar technical procedures, particularly for the determination 
 of venous pressure in the torcular Herophili. The very wide divergences 
 in these normal pressures (13 to 601 mm. in Becht's series) indicate 
 that the method is subject to many experimental defects. Weed and 
 Hughson (74), with these technical disadvantages in mind, devised a 
 simple method for recording intracranial venous pressure in the superior 
 sagittal sinus as it empties into the torcular. The procedure possessed 
 the very distinct advantage of permitting direct observations of the 
 effect of the manipulative procedure upon the pressure of the cerebro- 
 spinal fluid, thus affording control of artificial increases in the pressure 
 of the cerebrospinal fluid due to venous obstruction in the cranium. 
 With such technical controls, Weed and Hughson were able to show that 
 in practically every case the pressure of the cerebrospinal fluid was 
 considerably above (5-50 mm.) that of the sagittal sinus. They also 
 presented data which indicated, in accordance with the findings of 
 Dixon and Halliburton and of Becht, that alteration in intracranial 
 venous pressure effected alterations in the pressure of the cerebrospinal 
 fluid in the same direction but of lesser magnitude. Conversely it 
 was shown, in agreement with Dixon and Halliburton, that within the 
 physiological limits tested, alteration in the pressure of the cerebrospinal 
 fluid caused changes of lesser extent but of the same direction in the 
 sagittal venous pressure. 
 
 This conception, advanced by Weed and Hughson, that the pressure 
 of the cerebrospinal fluid is practically always above that of the cerebral 
 veins, alone affords basis of explanation for Wegefarth's experiments. 
 Wegefarth (77) demonstrated that an experimental communication 
 between subarachnoid space and superior sagittal sinus remained patent, 
 without hemorrhage into the meningeal cavities, for at least 4 days. 
 Removal of cerebrospinal fluid in these animals, however, resulted in 
 
190 LEWIS H. WEED 
 
 immediate intrameningeal hemorrhage. Such an observation, free from 
 the errors of recording instruments, can be explained only on 'the basis 
 that normally the pressure of the fluid is above that in the sinus or that 
 the former is constantly being reduced toward the latter. 
 
 Analysis of the reliable data concerning these normal relationships 
 seems convincing in demonstrating that the pressure of the cerebrospinal 
 fluid practically always exceeds that of the superior sagittal sinus. In 
 no sense may the two pressures be regarded as identical, for such an 
 identity of pressures is found only in animals in which the technical 
 procedures have resulted in the production of direct communications 
 between sinus and meningeal spaces. In the normal animal intracranial 
 arterial pressure is a factor of importance in the maintenance of the 
 pressure of the cerebrospinal fluid, though slight or slowly effected 
 changes in this arterial tension have but little influence upon the fluid 
 pressure. Thus while dependent upon both intracranial arterial and 
 venous pressures and while influenced passively and in the same direc- 
 tion by both, the pressure of the cerebrospinal fluid may be considered 
 to be relatively independent of both in that normally it maintains an 
 individual, fairly constant level far below that of the intracranial 
 arteries and somewhat above that of the intracranial veins. 
 
 It becomes desirable, then, to ascertain the factors which determine 
 the level of the cerebrospinal fluid pressure, but in this inquiry there are 
 practically no data available and the problem becomes speculative. Yet 
 it is instructive to think of the cerebrospinal fluid as being largely elabo- 
 rated by the cells of the choroid plexuses where the pressure in the 
 blood capillaries is estimated at from 40 to 60 mm. Hg. On the outer 
 side of these cells is the cerebrospinal fluid with its pressure of 110 to 
 130 mm. of Ringer's solution. After circulating through the cerebral 
 ventricles and subarachnoid space this fluid is largely returned into the 
 venous sinuses of the dura where the pressure (as determined in the 
 superior sagittal sinus) is below that of the cerebrospinal fluid (as de- 
 termined in the cisterna magna) . On such a basis the normal mecha- 
 nism for the absorption of this characteristic body liquid may well be a 
 simple process of filtration though the factors of osmosis and diffusion 
 are not excluded in the passage of the fluid through the cell-membrane 
 of the arachnoid villus. No determinations of the pressure of the cere- 
 brospinal fluid in the arachnoid villus have been made but it is unlikely 
 that this pressure is markedly different from that of the cisterna magna. 
 Obstruction to any part of the pathway of the fluid results in raising 
 intraventricular tension (cf. Nanagas (53)) thus demonstrating that 
 
THE CEREBROSPINAL FLUID 191 
 
 cerebrospinal fluid can be produced against higher pressures than nor- 
 mally exist in the cerebral ventricles. The normal pressure therefore 
 may be largely determined by the balance between the constant new 
 production within the cerebral ventricles and the absorption into the 
 dural sinuses : this pressure of the cerebrospinal fluid becomes dependent 
 also upon intracranial arterial and venous pressures, not only because 
 of the relation of these latter pressures to the production and absorption 
 of the fluid, but because of the constancy of volume of the intracranial 
 contents. 
 
 MODIFICATION OF THE PRESSURE OF THE CEREBROSPINAL FLUID 
 
 During the past ten years a number of investigators has studied the 
 alterations in pressure of the cerebrospinal fluid effected by the intro- 
 duction of various substances into the blood stream or alimentary canal. 
 The subject-matter naturally differentiates itself from the purely 
 mechanical modifications effected by pressure-changes in the cerebral 
 blood vessels. The general findings in this latter group of experiments 
 have been presented in the foregoing section of this review ; the pressure- 
 changes effected in the cerebrospinal fluid by solutions of various con- 
 centrations and by certain pharmacological agents and tissue extracts 
 will be discussed here. 
 
 Solutions of various concentrations. In 1919 Weed and McKibben (75) 
 reported that the pressure of the cerebrospinal fluid could be markedly 
 altered by the intravenous injection of solutions of various concentra- 
 tions. It was shown that such administration of strongly hypertonic 
 solutions lowered the pressure of this liquid to an extreme degree, fre- 
 quently producing negative values; with hypotonic solutions (distilled 
 water) a prolonged rise in the pressure of the fluid was obtained. 
 Ringer's solution in large doses produced a temporary increase in the 
 pressure of the cerebrospinal fluid, followed quickly by a return to nor- 
 mal levels. Accompanying these changes in fluid pressure, Weed and 
 McKibben (76) found marked alterations in the volume of the brain, 
 the hypertonic solution producing a small shrunken brain while the 
 hypotonic solution caused an outspoken swelling of the brain-substance. 
 The experimental changes in brain volume were particularly pro- 
 nounced in animals in which the cranial cavity had been opened by 
 trephining. 
 
 These physiological findings have since been abundantly confirmed 
 and clinical applications of the phenomena have been developed. 
 Gushing and Foley (15) demonstrated that similar alterations in the 
 
192 LEWIS H. WEED 
 
 pressure of the cerebrospinal fluid could be brought about by the inges- 
 tion of hypertonfc and hypotonic solutions. Subsequently Foley and 
 Putnam (31), after verifying the original conclusions, studied similar 
 changes in the pressure of the cerebrospinal fluid which were effected 
 by the intra-intestinal administration of these solutions of various 
 concentrations. And Sachs and Malone (60) reported observations 
 upon the decrease of brain volume caused by the intravenous injection 
 of strongly hypertonic solutions. In addition to these papers, there has 
 appeared a number of reports of clinical application of this experimental 
 modification of fluid pressure or brain bulk notably those of Gushing 
 and Foley (15), Sachs and Belcher (59), Ebaugh and Stevenson (22), 
 Foley (30) and Hughson (40). 
 
 Recently Weed and Hughson (72), (74) have extended the original 
 observations of Weed and McKibben (75); in addition to confirming 
 the initial work in detail, they have presented data showing the general 
 systemic and intracranial vascular alterations effected by these agents. 
 These observations were made over periods of from 2 to 7 hours, under 
 adequate experimental conditions in which the pressures recorded (cere- 
 brospinal fluid, carotid artery, superior sagittal sinus, superficial bra- 
 chial vein) remained surprisingly constant. The intravenous injection 
 of a large quantity of Ringer's solution was shown to cause a temporary 
 rise in the pressure of the cerebrospinal fluid with increases in both 
 sagittal and brachial venous pressures, the former being the greater. 
 At the end of the injection-period all of these pressures tended to return 
 to their previous levels, normal pressures customarily being attained 
 within 30 minutes. After the intravenous injection of distilled water 
 in similar amount, a prolonged rise in the pressure of the cerebrospinal 
 fluid occurred, accompanied by alterations in both sagittal and brachial 
 pressures. Both of these venous pressures increased during the period 
 of injection and for a few minutes thereafter, the sagittal outstripping 
 the brachial; within 30 minutes these pressures were usually returned 
 to their pre-injection levels, though the pressure of the cerebrospinal 
 fluid was still elevated. With the intravenous injection of strongly 
 hypertonic solutions, the pressure-alterations were most striking, the 
 cerebrospinal fluid, after a frequently occurring rise during the interval 
 of injection, dropping markedly and often exhibiting extreme negative 
 values. The pressure in the superficial brachial vein rose during the 
 period of the hypertonic injection and then rapidly resumed its pre- 
 injection level or a new level slightly below. More significant were the 
 alterations in sagittal venous pressure: here the reaction during the 
 
THE CEREBROSPINAL FLUID 193 
 
 period of injection depended largely on the reaction of .the pressure of the 
 cerebrospinal fluid, but following this the sagittal pressure was always 
 lowered to a greater extent than was the brachial venous pressure. 
 
 These experiments afforded a unique opportunity for study of the 
 mechanisms which normally control the pressure of the cerebrospinal 
 fluid. Analysis of the data demonstrated that alterations in the pres- 
 sure of the cerebrospinal fluid could be effected and maintained indepen- 
 dently of change in the intracranial and systemic vascular pressures. 
 The most striking similarities in reactions were those between the pres- 
 sure of the cerebrospinal fluid and that of the brachial vein and sagittal 
 sinus; after the injection of Ringer's solution or of distilled water they 
 exhibited somewhat the same alterations, differing not only in magni- 
 tude but in duration. After the intravenous injection of hypertonic 
 solutions, however, the relationships of the pressures were markedly 
 altered, with the pressure of the cerebrospinal fluid profoundly lowered, 
 the sagittal venous considerably and the brachial venous but little if 
 at all. Both brachial and sagittal venous pressures were found to be 
 lower than the pressure of the cerebrospinal fluid in the control- 
 periods; this relationship held after the injection of isotonic and hypo- 
 tonic solutions but was reversed when hypertonic solutions were given. 
 Arterial changes, when slowly brought about, caused little if any altera- 
 tion in the pressure of the cerebrospinal fluid, but when abrupt, their 
 influence was marked. 
 
 The changes in the pressure of the cerebrospinal fluid, effected by 
 the intravenous injection of solutions of various concentrations, must in 
 the final analysis find their explanation in the alteration of the osmotic 
 pressure of the circulating blood. The injection of a large volume of an 
 isotonic solution was followed by a short-enduring rise of the cerebro- 
 spinal fluid pressure which subsided in approximately the same time- 
 interval required for the pressure-changes effected by the hypertonic 
 and hypotonic solutions to reach their maxima. The usual time for 
 maximal reaction of the pressure of the cerebrospinal fluid was noted 
 to be from 25 to 35 minutes after the end of the intravenous injection; 
 in this interval the organism was attempting to compensate for altera- 
 tion in the volume and salt-content of the blood. With the isotonic 
 solutions, the compensation was one for increased volume of fluid only; 
 this compensation, if judged by the time of return of the cerebrospinal 
 fluid pressure to normal, was rapidly achieved. When, however, not 
 only the volume of the circulating blood was increased but its salt-con- 
 tent relatively diminished as by the intravenous injection of distilled 
 
194 LEWIS H. WEED 
 
 water, two processes of adjustment proceeded. The blood tended to 
 reestablish its normal salt-content by passage of water into the tissues 
 and possibly into some of the body-fluids, and by attraction of salts 
 from these places; and it also tended to compensate further by altera- 
 tion of the vascular bed. The increase in the pressure of the cerebro- 
 spinal fluid and in the brain volume may be taken to be a rough index 
 of the passage of fluid from blood vessel to tissue; the return of the 
 vascular pressures to normal levels while the pressure of the cerebro- 
 spinal fluid remained high, indicated the completion of certain of the 
 phases of readjustment. In the readjustments effected by the organism 
 to the injection of hypertonic solutions, there were somewhat similar 
 phases, yet differing because the great increase of fluid-volume in the 
 circulating blood was not immediate but was due to the attraction of 
 water from the body-tissues and possibly from the body-fluids a phe- 
 nomenon shown by the decrease in brain volume and by the reduction 
 of the pressure of the cerebrospinal fluid. 
 
 Such an explanation of the phenomena reported leads one naturally 
 to a consideration of the role played by the osmotic pressure of the blood 
 in the normal process of absorption of the cerebrospinal fluid. And 
 intimately connected with such a problem is that of the volume of the 
 brain in its relation to the intracranial pressure. At the present time 
 there are available no data which will permit of exact statement; it 
 must be realized that the osmotic changes in the blood, effected by such 
 relatively large injections of solutions of various concentrations, are 
 probably beyond the ordinary physiological limits of change. But if 
 one may judge merely by the anatomical and physiological evidence 
 afforded by subarachnoid injection of isotonic foreign solutions, osmosis 
 and diffusion play subordinate roles in the normal process of absorption 
 of the cerebrospinal fluid into the blood stream. 
 
 Pharmacological agents and tissue extracts. Most of the work done on 
 this subject has been actuated by the tenet that alteration in the pres- 
 sure of the cerebrospinal fluid, without significant change in intracranial 
 vascular pressures, affords a more reliable means of determining the 
 effect of these various agents upon the rate of production of the fluid 
 than the outflow method. While logically this subject should perhaps 
 be discussed under the heading of the modification of the rate of elabora- 
 tion of the liquid, it may properly be treated here as an experimental 
 alteration of the fluid pressure. 
 
 It may be stated at the outset that there is no unanimity of opinion 
 regarding the effect of either pharmacological agents or tissue extracts 
 
THE CEREBROSPINAL FLUID 195 
 
 upon the pressure of the cerebrospinal fluid, without significant altera- 
 tion in intracranial arterial or venous pressures. Dixon and Halli- 
 burton (20) found that extracts of the choroid plexus, chloroform, ether, 
 urethane, carbon dioxide, amyl nitrite, pilocarpine and other drugs 
 caused a "secretory rise" in cerebrospinal fluid pressure which was 
 independent of the intracranial vascular alterations. Becht (1) has 
 investigated the subject from the same angle, using methods somewhat 
 similar, and has stated that (p. 124) "all the changes in the fluid pressure 
 and in the fluid outflow which have been offered as proof of the secretory 
 mechanism of formation of the cerebrospinal fluids can be traced to 
 alterations in venous and arterial pressures in the skull." Similar con- 
 clusions have been reached by Becht and Matill (3) and by Becht and 
 Gunnar (2). 
 
 With this conflicting evidence it is of course impossible to do other 
 than reserve opinion in the matter. But certain aspects of the con- 
 troversy may be commented upon. Dixon and Halliburton and Becht 
 determined cerebral venous pressures in the torcular Herophili a meth- 
 od which because of the wide divergences in normal pressures reported 
 does not seem adequate though qualitative changes in the pressures are 
 probably fairly accurately shown. While the simple manometric meth- 
 od has been modified by Becht and Gunnar (2), it still has certain 
 limitations in determining a true change in rate of production of cere- 
 brospinal fluid. Of these, the fact that the normal channels of absorp- 
 tion are intact and functioning is the most obvious though Becht has 
 minimized the weight of this objection. There is however a much 
 more formidable disadvantage in that the manometric method cannot 
 take into account the experimental alteration of the volume of the brain. 
 The pressure-changes in the cerebrospinal fluid, effected by the intra- 
 venous injection of distilled water, appear to fulfil, after the period of 
 acute vascular change, all of the conditions necessary for the determina- 
 tion that the injection has caused an increased production of fluid a 
 markedly elevated cerebrospinal fluid pressure with intracranial vascular 
 pressures at the pre-injection levels. The outspoken increase in brain 
 volume under such conditions, however, may well be the sole explana- 
 tion of the phenomenon; until the experimental variations in brain 
 volume are more fully understood, the evidence obtained by the mano- 
 metric method should be accepted only with reservations. 
 
196 LEWIS H. WEED 
 
 MODIFICATION OP RATE OF OUTFLOW OF CEREBROSPINAL FLUID 
 
 That certain pharmacological substances may modify the rate of 
 flow of cerebrospinal fluid from a cannula introduced into the subarach- 
 noid space was first determined by Cappelletti (9) in 1900. Employ- 
 ing Cavazzani's (11) method of making a fistula into the cistern region, 
 Cappelletti showed that ether given by intratracheal tube in a curarized 
 dog increased the rate>of flow of the fluid from 0.15 to 0.35 grams per 15- 
 minute interval to 4.72 grams for the same interval. A second, a third 
 and a fourth administration gave momentary increases but not of the 
 same extent as the initial. Similar positive results were obtained with 
 pilocarpine, but the increases though obvious were not marked. Very 
 slight augmentation of the rate of outflow was also obtained with amyl 
 nitrite, while atropine and hyoscyamine caused a decrease and on repeti- 
 tion a cessation of the outflow. 
 
 Pettit and Girard (55) immediately confirmed and extended these 
 experimental findings of Cappelletti, including in their studies histologi- 
 cal examination of the choroid plexuses. And Meek's (49) observations 
 were likewise entirely confirmatory. The scope of the investigation 
 was widened in 1913 by Dixon and Halliburton (19) who studied the 
 effect of a large number of substances upon the rate of outflow of cere- 
 brospinal fluid from an occipito-atlantoid cannula. They were able to 
 classify the substances into four groups according to their effect on this 
 rate of outflow, placing the volatile anesthetics, alcohol, carbon dioxide 
 and extracts of choroid plexus and of brain in the group which caused 
 marked increase in secretion. Slight increases in the outflow were found 
 to be caused by large injections of water or of normal saline, cholesterin, 
 kephalin, atropine, pilocarpine and amyl nitrite. In the large third 
 group of substances which caused no increase or a diminution of secre- 
 tion were included extracts of the pituitary, of mussel, of pineal and of 
 pia mater, glucose, urea, lecithin, etc., while in the last group where the 
 effect was possibly masked by vascular or respiratory changes, were 
 muscarin, pilocarpine, adrenalin, etc. 
 
 Shortly thereafter Dandy and Blackfan (18), obtaining cerebrospinal 
 fluid by introduction of a special cannula through the atlas, found 
 marked accelerations of the rate of output of cerebrospinal fluid follow- 
 ing administration of ether and slight augmentations after pilocarpine. 
 With amyl nitrite and extracts of choroid plexus and of posterior lobe 
 of the hypophysis, no change in the rate of output of the fluid was 
 observed. 
 
THE CEREBROSPINAL FLUID 197 
 
 Realizing the limitations of this technique in that the normal channels 
 of absorption were intact and that the intracranial pressure was reduced 
 to the resistance of the needle, Weed and Gushing (71) in 1915 cathete- 
 rized the third ventricle and studied the outflow from the catheter 
 whose resistance was established at approximately normal pressure of the 
 fluid. In addition, cerebrospinal fluid was obtained by callosal and oc- 
 cipito-atlantoid punctures with needles of similarly standardized resist- 
 ances. Under these circumstances the intravenous injection of extract of 
 posterior lobe of the hypophysis was found to increase the outflow of 
 cerebrospinal fluid. This finding was explained by Dixon and Halliburton 
 (21) on the basis that the hypophysial extract had caused a contraction 
 of the bronchioles and consequent asphyxia. Dixon and Halliburton 
 used an intermittent blast for their artificial respiration while Weed and 
 Gushing employed intratracheal insufflation: it seems questionable 
 whether this explanation of the finding will suffice. 
 
 At this time also, Frazier and Peet (34) reported that brain-extract 
 increased the secretion of the cerebrospinal fluid as determined by out- 
 flow and that thyroid extract decreased it, independently of any vascular 
 changes. 
 
 The later studies of the effect of these substances upon the rate of 
 production of cerebrospinal fluid have been made by the manometric 
 method and have been discussed in the preceding section of this review. 
 The limitations of the outflow method were realized by Weed and Gush- 
 ing (71) in 1915; their modifications introduced control for some of the 
 sources of error but were incomplete. As Becht (1) has pointed out, 
 practically all of this work is of no scientific value because of failures to 
 record simultaneously the intracranial arterial and venous pressures. 
 Using this standard but employing the manometric method, Becht and 
 Matill (3) have concluded that there is no indisputable evidence that 
 the tissue extracts tested have a specific action on the cerebrospinal 
 fluid. And recently Becht and Gunnar (2) reported that adrenalin, 
 pituitrin, pilocarpine and atropine did not increase the production of 
 cerebrospinal fluid, as determined by manometer readings. It is true 
 that the method of recording the rate of outflow of cerebrospinal fluid 
 from a cistern cannula, even with careful determinations of intracranial 
 vascular pressures, yields unreliable data, but in many ways, also, the 
 manometric method fails. Both of these methods, which at the present 
 time are the only technical approaches to the problem, are of question- 
 able value because they both fail to take account of the experimental 
 variation in brain-bulk. When a method which will permit of actual 
 
198 LEWIS H. WEED 
 
 determination of this variable brain bulk, with observations also of 
 cerebrospinal fluid pressure and with a record of intracranial vascular 
 pressures, is devised, data of conclusive value will be obtained. And 
 yet one cannot but lay stress upon the changes in the choroidal epithe- 
 lium recorded by Pettit and Girard (55) after the injection of pilocarpine. 
 Likewise, as first reported by Cappelletti (9) and since noted by many 
 workers, the rapidly decreasing responses to ether and pilocarpine sug- 
 gest strongly that the accelerations of flow of cerebrospinal fluid under 
 these conditions were not due solely to vascular alteration, for such 
 ready fatigability would not seem to be associated with a vasomotor 
 reaction. 
 
 In this connection it is interesting to speculate upon the possibility 
 of modification of the rate of elaboration of the cerebrospinal fluid, after 
 the intravenous injection of solutions of various concentrations. There 
 is as yet no evidence of value in this regard, though Foley and Putnam 
 presented data which suggested that after the injection of a strongly 
 hypertonic solution, a new ratio between the rate of production and 
 absorption of the cerebrospinal fluid became established. But the final 
 elucidation of this phase of the problem will require additional work 
 before definite conceptions are acquired. 
 
 EELATIONSHIP OF CEREBROSPINAL FLUID TO NERVOUS SYSTEM 
 
 Many phases of the relationship existing between the central nervous 
 system and the cerebrospinal fluid are of utmost significance in the 
 present discussion. Filling the cerebral ventricles and central canal of 
 the spinal cord, the fluid also completely surrounds the cerebrospinal 
 axis in the subarachnoid space. This double relationship has prompted 
 many observers to look upon the cerebrospinal fluid as constituting a 
 fluid-cushion for the central nervous system within the closed system of 
 cranium and vertebral column. It has also prompted other workers to 
 liken the cerebrospinal fluid to the lymph of the nervous system a 
 conception which in the light of present knowledge of the lymphatic 
 system is untenable. 
 
 Halliburton (37), in a recent lecture, declared that the cerebrospinal 
 fluid serves as the lymph of the brain, though clearly differentiating it 
 from the true lymph of the lymphatic vessels. It seems likely that such 
 a designation, even when correctly qualified as Halliburton has stated 
 it, is apt to introduce error. All modern investigators of the lymphatic 
 system are agreed that true lymphatic vessels do not exist within the 
 dura mater; the older descriptions of such lymphatic vessels were actu- 
 
THE CEREBROSPINAL FLUID 199 
 
 ally descriptions of intradural tissue-channels, subpial tissue-channels, 
 or arachnoidal cell-columns within the dura mater. As Sabin (58) has 
 pointed out, our knowledge of the lymphatic system has advanced so 
 that it becomes now necessary to restrict the term "lymph" to the fluid 
 contained within true lymphatic vessels and not to use it to designate 
 any body-fluid. 
 
 But in one respect the cerebrospinal fluid does function as an acces- 
 sory fluid to the central nervous system. In foregoing sections the 
 drainage of the fluid contained within the perivascular channels toward 
 the subarachnoid space has been commented upon; this fluid really 
 becomes added to the ventricular cerebrospinal fluid in the subarach- 
 noid space. In that sense, then, these perivascular spaces represent 
 accessory drainage channels, uninterrupted by cell-membranes and 
 capable of carrying toward the subarachnoid space the waste products of 
 nerve-cell activity. Lacking a true lymphatic system, the nervous 
 tissue apparently makes use of these perivascular channels as pathways 
 for fluid elimination. 
 
 The ultimate connection of these perivascular channels with potential 
 spaces about each nerve-cell indicate the close relationship between the 
 cerebrospinal fluid and the nervous system. And in addition to these 
 rather obvious fluid spaces about the nerve-cells, there is evidence 
 indicating that this fluid-system is intimately connected with the general 
 tissue-channels through the ground-substance of the brain. The 
 general direction of flow of this fluid under normal conditions seems to 
 be toward the subarachnoid space. 
 
 But under certain conditions this direction of flow may be reversed 
 so that the cerebrospinal fluid passes from subarachnoid space to nerve- 
 cell. The first of these conditions is that of cerebral anemia in which, 
 as Mott (52) showed by histological study, all of the perivascular, 
 pericapillary and perineuronal spaces are dilated. The author (68) 
 made use of this phenomenon as a means of injecting this perivascular 
 system from the subarachnoid space. The second of these conditions 
 under which the perivascular flow is toward nerve-cell, is brought about 
 by the intravenous injection of strongly hypertonic solutions. This 
 phenomenon was first noted by Weed and McKibben (76) who supplied 
 a foreign solution of sodium ferrocyanide and iron-ammonium citrate 
 to the subarachnoid space at the time when the cerebrospinal fluid 
 pressure was approaching zero, following the intravenous injection of a 
 strongly hypertonic solution. This foreign solution was subsequently 
 found (p. 536) "to have passed from the subarachnoid space along the 
 
200 LEWIS H. WEED 
 
 perivasculars into the substance of the nervous system, reaching the 
 interfibrous spaces in the white matter and the pericellular spaces in 
 the gray." These observations were interpreted as indicating that, 
 under the influence of the intravenous injection of the strongly hyper- 
 tonic solution, the dislocation of a considerable quantity of cerebro- 
 spinal fluid into the nervous system occurred. 
 
 Foley (29) has subsequently carried out experiments quite similar to 
 those reported by Weed and McKibben, using the same foreign salts 
 for subarachnoid introduction and intravenous injections of strongly 
 hypertonic solutions. In addition to the findings already detailed, 
 Foley obtained evidence of a retrograde absorption not only by epen- 
 dyma but by choroid plexuses. The absorption by the ependyma is 
 amply verified by the work of Wislocki and Putnam (79) and Nanagas 
 (53), but these latter workers, using careful histological control, have 
 been unable to obtain evidence of absorption of the foreign salts by 
 the choroid plexuses. 
 
 And in work as yet unpublished the writer has repeated many of his 
 earlier experiments done with McKibben, with findings confirmatory in 
 every regard. The intracranial vascular and the cerebrospinal fluid 
 pressures have been determined both before and throughout the period 
 of subarachnoid introduction of the foreign solution, so that definite 
 physiological control is afforded. The results indicate that with the 
 increase of osmotic pressure of the blood, due to the intravenous injec- 
 tion of hypertonic solutions, the cerebrospinal fluid is aspirated into the 
 shrinking nervous system, chiefly along the perivascular channels but 
 also through the ependymal lining of the ventricles. Along these chan- 
 nels, under this extraordinary osmotic pull, actual absorption of 
 the fluid into the vessels of the nervous tissue takes place. The 
 findings suggest a reversal, following the injection of the hypertonic 
 solution, of the normal processes; the osmotic pressure of the blood 
 stream, under these conditions, seems to be a determining factor in the 
 absorption of the cerebrospinal fluid. Interpretation of certain of the 
 experimental observations makes it seem likely that diffusion also plays 
 a part in the process. 
 
 RESUME 
 
 The limitations of this review have made it impossible to more than 
 rather briefly discuss a few of the many problems connected with the 
 cerebrospinal fluid. Many equally absorbing phases have been 
 untouched for one or another reason but an attempt has been made to 
 
THE CEREBROSPINAL FLUID 201 
 
 indicate the type of evidence which has furnished the working 'hypotheses 
 and to point out the limitations of the procedures on which so many 
 conclusions have been based. 
 
 Our present knowledge of the processes of the cerebrospinal fluid in 
 many respects is inadequate. The conception that this characteristic 
 body-fluid is largely produced by the intraventricular choroid plexuses 
 is based not on any single conclusive piece of evidence but on a 
 mass of suggestive data; when considered from all standpoints, however, 
 the hypothesis seems today well established. The current ideas regard- 
 ing the circulation of the fluid through cerebral ventricles and sub- 
 arachnoid space are founded largely on exact anatomical evidence, 
 particularly in regard to the structure of the meninges and the use of 
 these intrameningeal channels as fluid-pathways. And likewise, there 
 are firm and reliable data of an anatomical and physiological nature 
 supporting the contention that the cerebrospinal fluid is absorbed 
 largely into the venous system and to a lesser extent into the lymphatic 
 channels. It is possible now to discard the hypothesis of equality . 
 between the cerebrospinal fluid pressure and that of the cerebral veins, 
 and to regard the cerebrospinal fluid as being maintained at an indi- 
 vidual, relatively independent pressure at fairly constant levels above 
 that of the sagittal venous sinus. The conceptions of pressure-changes 
 effected by the intravenous injection of solutions of various concentra- 
 tions are substantiated by dependable observations, but it does not seem 
 as yet justifiable to accept, without further control, the data furnished 
 in regard to similar changes brought about by administration of phar- 
 macological agents and tissue extracts. And the same cautions may 
 be urged in regard to the acceptance of conclusions based on the effects 
 of various agents upon the rate of outflow of the fluid. 
 
 Yet these problems are but few of the many fascinating subjects of 
 investigation in this field. The interesting questions of the chemical 
 composition of the fluid have not been discussed: is the cerebrospinal 
 fluid a true secretion, a transudate, or a modified dialysate? Likewise, 
 the long-debated problems of the passage of foreign salts, of drugs, etc., 
 from blood stream into the fluid must be left for future review, though 
 with possibly a note of suggestion that these investigations be carried 
 out with control of the cerebrospinal fluid pressure. And so may the 
 many other partially answered questions centering about this fluid be 
 enumerated. 
 
 But in this field of research the work of the next few years will solve 
 certain problems; yet the solution of these will but expose wider fields 
 
202 LEWIS H. WEED 
 
 for examination. Here, as in countless other investigations, the study of 
 structure must proceed hand in hand with the study of "function, 
 for many of the erroneous conceptions introduced into the literature of 
 the cerebrospinal fluid have been due to failure to give regard to one or 
 other of these basic factors. Future investigations will be the more 
 profitable if the studies be largely along the lines of physiological- 
 anatomical control. 
 
 BIBLIOGRAPHY 
 
 (1) BECHT: Amer. Journ. Physiol., 1920, li, 1. 
 
 (2) BECHT AND GTJNNAR: Amer. Journ. Physiol., 1921, Ivi, 231. 
 
 (3) BECHT AND MATILL: Amer. Journ. Physiol., 1920, li, 126. 
 
 (4) BERGMANN: Deutsch. Chirurgie, 1880, xxx, 26,6. 
 
 (5) BERGMANN: Arch. f. Klin. Chirurgie, 1885, xxxii, 705. 
 
 (6) BLAKE: Journ. Comp. Neurol., 1900, x, 79. 
 
 (7) BURROWS: On disorders of the cerebral circulation, (London, 1846.) Phila- 
 
 delphia, 1848. 
 
 (8) CANNIETT: Journ. de Med. de Bordeaux, 1897, xxvii, 547. 
 
 (9) CAPPELLETTI: Arch. Ital. de Biol., 1900, xxxv, part 2, 463. 
 
 (10) CATHELIN: Le Circulation du Liquide Cephalo-Rachidien, Paris, 1912. 
 
 (11) CAVAZZANI: Atti d. Accad. d. Sci. Med. e Nat. di Ferrara, 1899, Ixxiii, 27. 
 
 (12) CLAISSE AND LEVY: Bulletin de la Societe anatomique, 1897, 265. 
 
 (13) GUSHING : v Mitt. a. d. Grenzgebieten d. Med. u. Chir., 1902, ix, 773. 
 
 (14) GUSHING: Journ. Med. Research, 1914, xxxi (N. S. xxvi), 1. 
 
 (15) GUSHING AND FOLEY: Proc. Soc. Exper. Biol. Med., 1920, xvii, 217. 
 
 (16) DANDYX Trans. Amer. Surg. Assoc., 1919, xxxvii, 397. 
 
 (17) DANDY AND BLACKFAN : Journ. Amer. Med. Assoc., 1913, Ixi, 2216. 
 
 (18) DANDY AND BLACKFAN : Amer. Journ. Dis. Children, 1914, viii, 406. 
 
 (19) DIXON AND HALLIBURTON: Journ. Physiol., 1913, xlvii, 215. 
 
 (20) DIXON AND HALLIBURTON: Journ. Physiol., 1914, xlviii, 128. 
 
 (21) DIXON AND HALLIBURTON: Journ. Physiol., 1916, 1, 198. 
 
 (22) EBAUGH AND STEVENSON: Johns Hopkins Hosp. Bull., 1920, xxxi, 440. 
 
 (23) ENGEL: Arch. f. Zellforschung, 1909, ii, 191. 
 
 (24) ESSICK: Contributions to Embryology no. 42, Publ. no. 272, Carnegie Inst. 
 
 Washington, 1920, 377. 
 
 (25) FAIVRE: These de Paris, 1853, (no. 142, dxl). 
 
 (26) FALKENHEIM AND NAUNYN: Arch, exper. Path. u. Pharm., 1887, xxii, 261. 
 
 (27) FELTON, HUSSEY AND BAYNE-JONES: Arch. Int. Med., 1917, xix, 1085. 
 
 (28) FINDLAY: Brain, 1899, xxii, 161. 
 
 (29) FOLEY: Arch. Neur. and Psych., 1921, v, 744. 
 
 (30) FOLEY: Surg., Gynec. andObstet., 1921, xxxiii, 126. 
 
 (31) FOLEY AND PUTNAM: Amer. Journ. Physiol., 1920, liii, 464. 
 
 (32) FRANCINI: Sper. Arch-, di Biol., 1907, Ixi, 415. 
 
 (33) FRAZIER AND PEET: Amer. Journ. Physiol., 1914, xxxv, 268. 
 
 (34) FRAZIER AND PEET: Amer. Journ. Physiol., 1915, xxxvi, 464. 
 
 (35) GALEOTTI: Riv. di Pat. Nerv. e Ment., 1897, xii, 480. 
 
 (36) GOLDMANN: Vitalfarbung am Zentralnervensystems, Berlin, 1913. 
 
 (37) HALLIBURTON: Proc. Roy. Soc. Med., 1916, x, (Section of Neurology), 1. 
 
THE CEREBROSPINAL FLUID 203 
 
 (38) HESS: Morphol. Jahrbuch., 1885, x, 578. 
 
 (39) HILL: Physiology and pathology of the cerebral circulation, London, 1896. 
 
 (40) HUGHSON: Journ. Amer. Med. Assoc., 1921, Ixxvii, 1859. 
 
 (41) HWOROSTUCHIN : Arch. f. mikr. Anat., 1911, Ixxvii, 232. 
 
 (42) KELLIE: Trans. Med. Chir. Soc., Edinburgh, 1824. 
 
 (43) KEY AND RETZIUS: Anatomic des Nervensystems und des Bindesgewebe, 
 
 Stockholm, 1876 
 
 (44) LEWANDOWSKY: Zeitschr. f. Klin. Medizin, 1900, xl, 480. 
 
 (45) LEYDEN: Arch. f. path. Anat. u. Physiol., 1866, xxxvii, 519. 
 
 (46) LOEPEB: Compt. rend. d. 1. Soc. d. biol., 1904, Ivi, 1010. 
 
 (47) LUSCHKA: Die Adergeflechte des menschlichen Gehirns, Berlin, 1855. 
 
 (48) MAGENDIE: Recherches sur le Liquide Cephalo-rachidien, Paris, 1825. 
 
 (49) MEEK: Journ. Comp. Neur. and Psych., 1907, xvii, 286. 
 
 (50) MESTREZAT: Le Liquide Cephalo-rachidien, Paris, 1912. 
 
 (51) MONRO: Observations on the structure and functions of the nervous system, 
 
 Edinburgh, 1783. 
 
 (52) MOTT: Lancet, 1910, part 2, 1 and 79. 
 
 (53) NANAGAS: Johns Hopkins Hosp. Bull., 1921, xxxii, 381. 
 
 (54) PELLIZZI: Folia Neuro-Biologica. Internat. Zentralorgan f. d. gesammt. 
 
 Biol. d. Nervensystems, 1911, v, 305. 
 
 (55) PETTIT AND GIRARD: Archiv. d'Anat. Mic., 1902, v, 213. 
 
 (56) QUINCKE: Arch. f. Anat. u. Physiol., (Du Bois Reymond) 1872, 153. 
 
 (57) REINER AND SCHNITZLER: Centralbl. f. Physiol., 1894, viii, 684. 
 
 (58) SABIN: Harvey Lectures, series ix, New York, 1915-16. 
 
 (59) SACHS AND BELCHER: Journ. Amer. Med. Assoc., 1920, Ixxv, 667. 
 
 (60) SACHS AND MALONE: Amer. Journ. Physiol., 1921, Iv, 277. 
 
 (61) SCHLAPFER: Zeigler's Beitr. z. allgem. Path. u. path. Anat., 1905, vii, 101. 
 
 (62) SICARD AND CESTAN: Bull, et Mem. Soc. Med. d'Hop. de Paris, 1904, third 
 
 series, xxi, 715. 
 
 (63) SPINA: Arch. f. d. gesammt. Physiol., 1900-1901, Ixxxiii, 120 and 415. 
 
 (64) STUDNICKA: Anat. Hefte, 1900, xv, 303. 
 
 (65) TESTUT: Traite d'Anatomie Humaine, 1905, ii. 
 
 (66) WEED: Journ. Med. Research, 1914, xxxi, (N. S. xxvi), 21. 
 
 (67) WEED: Journ. Med. Research, 1914, xxxi, (N. S. xxvi), 51 
 
 (68) WEED: Journ. Med. Research, 1914, xxxi, (N. S. xxvi), 93. 
 
 (69) WEED: Contributions to Embryology no. 14, Publ. no. 225, Carnegie Inst. 
 
 Washington, 1917, 1. 
 
 (70) WEED: Anat. Record, 1917, xii, 461. 
 
 (71) WEED AND GUSHING: Amer. Journ. Physiol., 1915, xxxvi, 77. 
 
 (72) WEED AND HUGHSON: Amer. Journ. Physiol., 1921, Iviii, 53. 
 
 (73) WEED AND HUGHSON: Amer. Journ. Physiol., 1921, Iviii, 85. 
 
 (74) WEED AND HUGHSON: Amer. Journ. Physiol., 1921, Iviii, 130. 
 
 (75) WEED AND McKiBBEN: Amer. Journ. Physiol., 1919, xlviii, 512. 
 
 (76) WEED AND MCKIBBEN: Amer. Journ. Physiol., 1919, xlviii, 531. 
 
 (77) WEGEFARTH: Journ. Med. Research, 1914, xxxi, (N. S. xxvi), 149. 
 
 (78) WILDER: Journ. Nerv. and Ment. Dis., 1886, xiii, 206. 
 
 (79) WISLOCKI AND PUTNAM: Amer. Journ. Anat., 1921, xxix, 313. 
 
 (80) YOSHIMURA: Arbeiten a. d. neurol. Inst. a. d. Wien., 1910, xviii, 1. 
 
 (81) ZEIGLER: Arch. f. Klin. Chir., 1896, liii, 75. 
 
YD 2V255 
 
 MOLOGY LIBRARY