At least seven compounds synthesized by cultured cells in amounts which should suffice for sustained growth have nevertheless proved under certain conditions necessary for survival (asparagine, cystine, glutamine, homocystine inositol, pyruvate, serine). In every instance so far examined, that requirement has been population-dependent, disappearing at cell densities sufficiently large to bring the concentration in the medium and in the cellular pool to metabolically effective levels before the cells died of the specific deficiency. At population densities of less than 100 cells per ml, serine was required by all cultured cells so far studied. With more exigent strains, such as the RT6 strain of rabbit fibroblast and the P388 mouse leukemia, the serine requirement disappeared only at populations in excess of 50,000 and 150,000 cells per ml, respectively. The requirement for pyruvate by the latter cell as an alternative to serine also disappeared at that population density. In a cystine-free medium there were population-dependent requirements for cystine, homocystine, or serine, depending on the experimental conditions. With methionine and glucose as cystine precursors, the critical population density permitting cellular survival and growth was in excess of 200,000 cells/ml. The provision of homocystine as an intermediate reduced the critical population density to 10,000 to 60,000 cells/ml; with the further provision of serine, the critical cell concentration permitting growth was reduced to 50 to 500 cells/ml. Cells adapted to glutamic acid, and capable of utilizing it as a substitute for glutamine, nevertheless required exogeneous glutamine at cellular densities of less than 50,000 cells per ml. In some experiments, the provision of asparagine reduced the critical population density to 10,000 cells/ml, presumably because of its glutamine-sparing action. Inositol is required by most cell lines, despite their demonstrated capacity to synthesize it from glucose. With at least one cell line (HeLa), sustained growth was occasionally achieved in an inositol-free medium if the population density was maintained in excess of 240,000 cells/ml. The possible implications of these findings with respect to the loss of specific organ functions in dispersed cell culture are discussed.

A simplified technique has been described for the continuous growth of mammalian cells in suspension culture. The cell population density increased as the rate of input of fresh medium was decreased, and the average generation time was concommittantly prolonged. At relatively high input rates, the population remained stabilized for an indefinite period, but at low flow rates, there was sometimes a cyclical variation in population density. The factor limiting growth rate at input rates of approximately 0.2 volumes per day was not the exhaustion of the medium; but in some experiments a non-dialyzable material appeared which inhibited cell growth.

Monkey kidney cells tested in their first culture passage, 24 hours after their isolation from the animal host, required the same 13 amino acids for survival and growth as cell lines serially propagated in culture for years. Under the conditions of the present experiments, arginine, cystine, glutamine, histidine, and tyrosine proved necessary, over and above the 8 amino acids required for nitrogen balance in man. With the serially propagated lines, glutamic acid substituted for glutamine only at extremely high and non-physiological levels. In the monkey kidney cell cultures, however, glutamic acid and glutamine were interchangeable, mole for mole; and aspartic acid and asparagine were also effective as glutamine substitutes. Glycine was growth-stimulatory for monkey kidney cells in primary culture, and the cells grown in a glycine-deficient medium usually failed to survive subculture.

Only minimal amounts of poliomyelitis virus were formed by HeLa cells placed in a medium free from glucose and glutamine, even if the medium contained an otherwise full complement of essential and non-essential amino acids, purines, pyrimidines, NB 4 + , and serum protein. Conversely, within one log of the optimal yield of virus was formed by HeLa cells in a medium containing only glucose, glutamine, and salts, even if the cells had been starved in this medium for 12 hours prior to their inoculation. The presence of glucose alone caused an average 170-fold increase in viral output beyond the amounts formed by the glucose- and glutamine-depleted cells. The addition of glutamine alone caused an average 2000-fold increase; and the addition of both increased the viral formation 40,000-fold. Qualitatively similar results were obtained with unstarved cells, not previously depleted of glucose and glutamine. It follows that only a small proportion of HeLa cells are capable of forming virus unless either glucose or glutamine, or both, are present in the medium. The elaboration of virus was significantly delayed in media containing glucose but no glutamine. The absence of glucose and glutamine did not prevent the fixation of poliomyelitis virus by the cell. When these compounds were added to previously depleted cells even 6 hours after inoculation, and after the excess free virus had been removed by washing and by the addition of specific antiserum, normal amounts of virus were formed despite the degenerative changes caused by the previous glucose and glutamine deprivation. Possible functions of glucose and glutamine in the elaboration of virus are discussed in the text. Such factors other than glucose, glutamine, or salts ( e.g . amino acids, purines, pyrimidines, vitamins, protein, or NH 4 + ) as may be needed by HeLa cells for the propagation of poliomyelitis virus, need not be present in the medium and cannot be easily washed out of the cell. Even 12 hours' total deprivation of the cells in salt solution prior to inoculation only slightly decreased their virus-synthesizing capacity in a similarly deficient medium, provided only that adequate amounts of glucose and glutamine were retained.

Seven vitamins have to date proved essential for the survival and multiplication of a mouse fibroblast (strain L) and a human carcinoma cell (strain HeLa) in tissue culture: choline, folic acid, nicotinamide, pantothenic acid, pyridoxal, riboflavin, and thiamin. It was necessary to cultivate the cells for 5 to 15 days in a medium lacking the specific vitamin before the deficiency became apparent in the cessation of multiplication and the development of specific cytopathogenic effects. In their early stages these changes could be reversed by the addition of the missing vitamin, an in vitro , "cure" of a vitamin deficiency. The maximally effective concentrations were in the range 10 7 to 10 8 gm. per ml. The probability that additional vitamins not demonstrably essential under the conditions of the present experiments are nevertheless required for survival and growth is discussed in the text.

The amino acid requirements of a human uterine carcinoma cell (HeLa strain) have been defined. The 12 compounds previously found to be essential for the growth of a mouse fibroblast proved similarly essential for this human epithelial cell. They included arginine, cyst(e)ine, histidine, and tyrosine, in addition to the eight amino acids required for nitrogen balance in man (isoleucine, leucine, lysine, methionine, phenylalanine, threonine, tryptophan, and valine). Only the L -amino acids were active; the D -enantiomorphs had no demonstrable effect at physiologic concentrations. The minimum concentrations required for survival and limited growth varied from 0.003 µ M per ml. for L -tryptophan, to 0.1 µ M per ml. for L -lysine. The concentrations permitting optimum growth similarly varied from 0.01 µ M per ml. for tryptophan, to 0.1 µ M per ml. for leucine, isoleucine, threonine, lysine, and valine. The latter optimum concentrations of the individual amino acids were closely correlated with their serum levels. With at least six of the amino acids, high concentrations, in the range 1 to 10 µ M per ml., caused a definite growth inhibition. In the absence of a single essential amino acid, degenerative changes occurred in the cells, culminating in their death and dissolution. In the early stages, however, these degenerative changes could be reversed by the restoration of the missing component.

1. ( a ) Mammalian cells in tissue culture (mouse fibroblasts and malignant human uterine epithelium) did not concentrate penicillin from the culture medium. Even at low concentrations, the cellular accumulation was usually less than that in the surrounding fluid, and most of it was removed by washing. The radioactive material in such eluates was actively bactericidal, and was presumably in large part unchanged penicillin. ( b ) Penicilloic acid, produced by the action of penicillinase, was bound to the same (limited) extent as the active antibiotic. In both these respects mammalian cells behaved like naturally penicillin-resistant bacteria, and unlike such penicillin-sensitive bacteria as Streptococcus pyogenes or Diplococcus pneumoniae . 2. Cell-free sonic extracts of the L strain had the same limited reactivity with penicillin as the intact cells. The relatively minute amounts bound by the cells are therefore not due to their impermeability, but instead reflect the inherently low reactivity of the cellular constituents with penicillin. 3. It is suggested that the relative non-toxicity of penicillin for the mammalian host, and for mammalian cells in tissue culture, may be related to this low order of reactivity with the antibiotic.

1. In a previous study, the differing sensitivity of bacterial strains as they occur in nature appeared to be correlated with their correspondingly differing reactivity with penicillin. Presumably, the over-all reactivity of the cell with penicillin paralleled that of the vulnerable cell component(s). However, when penicillin-resistant variants of these strains ( Streptococcus pyogenes, Micrococcus pyogenes, Diplococcus pneumoniae , and Streptococcus faecalis ) were produced by serial passage through increasing concentrations of antibiotic, this correlation between resistance and the ability of the cell to bind penicillin was no longer apparent. Some resistant variants bound more penicillin than the parent, sensitive cell ( Streptococcus faecalis, Micrococcus pyogenes ), some were unchanged in their reactivity ( Diplococcus pneumoniae, Micrococcus pyogenes ), and some bound less ( Streptococcus pyogenes, Micrococcus pyogenes ). One resistant variant of Micrococcus pyogenes at first showed enhanced reactivity with penicillin; on continued passage through antibiotic, there was a further increase in resistance, but now associated with a significantly decreased reactivity. In the case of Diplococcus pneumoniae , a resistant variant at first reacted normally with penicillin; on continued passage in antibiotic, its binding affinity for penicillin gradually decreased, but with no associated further increase in resistance. 2. The reactivity with penicillin of cell-free sonic extracts of the resistant variants paralleled that of the intact organisms. Permeability considerations therefore did not seem involved in the increased resistance produced by serial passage in antibiotic. 3. The penicillin-resistant variants did not have an enhanced capacity to degrade the free intracellular antibiotic. 4. Possible alternative explanations are discussed in the text.

1. Bacteria exposed to C 14 - or S 35 -labelled penicillin bound and concentrated the antibiotic. The amount bound from low concentrations (0.001 to 0.01 µg. per ml.) was related to the penicillin sensitivity of the strain; and highly sensitive organisms such as Streptococcus pyogenes concentrated the antibiotic as much as 200-fold. The material bound by highly sensitive strains from these low concentrations of penicillin was removed to only a minor extent by washing. Penicillin inactivated with penicillinase, by hot H 2 SO 4 or by cold HCl was not similarly concentrated. 2. Despite wide differences in their sensitivity to penicillin, at equieffective (LD 99.9 ) levels, four of the five strains here studied had bound comparable amounts of antibiotic. That lethal intracellular concentration averaged 1.7 to 4 µg. per gm., and 1600 to 3300 molecules per cell. 3. Bacteria in the logarithmic phase of growth, "resting" organisms, and cell-free sonic extracts, had roughly the same reactivity with penicillin. The differences in the amount of penicillin bound by bacteria of varying sensitivity therefore do not rest on differences in the permeability of the cell. 4. Escherichia coli (K12)-inactivated penicillin within the cell; and that inactivation adequately explains its low binding affinity, and its relative resistance. The other species here studied inactivated the diffusible cellular penicillin to only a minor degree, and not enough to explain the differences in their penicillin-sensitivity. The working hypothesis is suggested that their varying sensitivity is instead determined by differences in the reactivity of vital cell components with the antibiotic. 5. At high concentrations of antibiotic (1 to 1000 µg. per ml.), there was non-specific binding by all the cell strains and cell extracts here studied, unrelated to their sensitivity, and masking the specific differences in reactivity noted at low concentrations. At these high levels, penicilloic acid was bound to the same degree as the active antibiotic.

The amount of penicillin which suffices to abort pneumococcal and streptococcal infections in white mice and rabbits, as well as syphilitic infection in rabbits, has been shown to increase markedly with the number of organisms inoculated. The curative dose increases also with the age of the infection, presumably owing to the interim increase in the number of organisms in the nfected host. The necessity for the larger doses of penicillin is attributed to the longer itime for which effectively bactericidal concentrations of penicillin must then be provided in order to kill the larger number of organisms.

1. The concentrations of penicillin G which ( a ) reduced the net rate of multiplication, ( b ) exerted a net bactericidal effect, and ( c ) killed the organisms at a maximal rate, have been defined for a total of 41 strains of α- and ß-hemolytic streptococci, Staphylococcus aureus and Staphylococcus albus, Diplococcus pneumoniae , and the Reiter treponoma. 2. The concentration which killed the organisms at a maximal rate was 2 to 20 times the minimal effective level ("sensitivity" as ordinarily defined). With some organisms, even a 32,000-fold increase beyond this maximally effective level did not further increase the rate of its bactericidal effect. However, with approximately half the strains here studied (all 4 strains of group B ß-hemolytic streptococci, 4 of 5 group C strains, 5 of 7 strains of Streptococcus fecalis , 2 of 4 other α-hemolytic streptococci, and 4 of 9 strains of staphylococci), when the concentration of penicillin was increased beyond that optimal level, the rate at which the organisms died was paradoxically reduced rather than increased, so that the maximal effect was obtained only within a relatively narrow optimal zone. 3. There were marked differences between bacterial species, and occasionally between different strains of the same species, not only with respect to the effective concentrations of penicillin, but also with respect to the maximal rate at which they could be killed by the drug in any concentration. Although there was a rough correlation between these two factors, there were many exceptions; individual strains affected only by high concentrations of penicillin might nevertheless be killed rapidly, while strains sensitive to minute concentrations might be killed only slowly. 4. Within the same bacterial suspension, individual organisms varied only to a minor degree with respect to the effective concentrations of penicillin. They varied strikingly, however, in their resistance to penicillin as measured by the times required to kill varying proportions of the cells.

The intravenous injection into rabbits of suspensions of dead T. pallidum derived from rabbit testicular chancres regularly caused the appearance of Wassermann and flocculation antibodies in significantly increased titer. Control suspensions of cultured treponemes (Reiter strain) added to extracts of normal testes were ineffective. This suggests that the Wassermann and flocculation reagin elaborated during syphilitic infection may be an antibody to T. pallidum which happens to cross-react with alcoholic extracts of mammalian tissue. The antisera did not cause the agglutination of suspensions of pathogenic T. pallidum , living or dead, did not give specific complement fixation with those suspensions, and did not usually cause the living treponemata to lose their infectiousness. Animals immunized with such aqueous suspensions for as long as 4 months, or with organisms suspended in a water-in-oil emulsion, were not demonstrably resistant to infection. As few as ten living organisms inoculated intradermally into animals "immunized" with as many as 38 billion dead treponemata regularly produced typical darkfield positive infections; and two of five animals inoculated intratesticularly with ten organisms were also infected. The contradiction involved in the production of antibodies cross-reacting with a non-specific antigen, and the non-appearance of specific antibodies against the organism used as antigen, is discussed in the text.

1. A relatively small amount of penicillin sufficed to abort syphilitic infection in rabbits when administered during the incubation period of the disease. 2. The abortive dose, given as a single intramuscular injection in oil and beeswax, varied with the age of the infection. ( a ) With a fixed intratesticular inoculum, the amount of penicillin necessary to prevent infection in half the animals remained at a constant level for 4 days. By the end of the 2nd week more than seven times this dosage was needed; and by the 6th week, after the chancre had appeared, more than thirty times the amount was needed to obtain the same result. The progressive increase in the abortive dose of penicillin with the passage of time probably reflects the interim multiplication of organisms. ( b ) Qualitatively similar results were obtained in rabbits inoculated intracutaneously. 3. The abortive dose varied also with the size of the inoculum. In animals inoculated intracutaneously with 20, 2,000, and 200,000 spirochetes, and treated 4 days later, it required 200, 500, and 3,500 units per kg., respectively, to protect half the animals, and the corresponding PD 90 dosages were 500, 2,000, and 16,000 units per kg. 4. The present observations, indicating the ease of aborting experimental rabbit syphilis during the incubation period by a single injection of penicillin, are perhaps applicable to the prevention of the disease in man. 5. Asymptomatic infections were rare in animals receiving inadequate doses of penicillin during the incubation period. Sixty-four of sixty-five such animals developed darkfield-positive lesions at the inoculated site. In animals treated 6 weeks after inoculation, however, after the development of lesions, inadequate treatment was usually manifested by an asymptomatic redissemination of organisms demonstrable only by lymph node transfer. The difference in the two groups probably reflects the beginning development of immunity in the rabbits treated 6 weeks after inoculation. 6. As suggested by Rake and Dunham (11), the treatment of animals during the incubation period permits a rapid assay of antisyphilitic agents, and one which requires only small amounts of material. Possible limitations of the method are discussed in the text.

1. Penicillins F, G, K, and X were all inactivated by human and rabbit serum, but two qualitatively distinct mechanisms were apparently involved. 2. One was a slow inactivation of all four penicillins by a relatively thermostable serum component which was not demonstrably affected by heating for 60 minutes at 56°C. ( a ) In both human and rabbit serum this general inactivation of penicillin behaved like a pseudo first order reaction, with a velocity constant of 0.05–0.07 for penicillin X, and 0.09–0.11 for penicillins F and G. ( b ) The percentage of penicillins F, G, and X inactivated per hour was independent of their concentration over the range 0.4 to 50 micrograms per cc., averaging 9.5, 10, and 6.5 per cent, respectively, in human serum, and 9,8.5, and 5 per cent in rabbit serum. ( c ) The rate of inactivation varied linearly with the concentration of the serum factor. ( d ) Penicillin X was consistently and significantly less susceptible to inactivation than any of the other penicillins. Although minor differences were observed between F and G, these were not consistent, and are of questionable significance. 3. Superimposed on this slow inactivation of penicillins F, G, K, and X by a thermostable serum component was a much faster inactivation observed only with penicillin K. ( a ) In both rabbit and human serum, the serum factor responsible for this inactivation was highly thermolabile, and was almost completely destroyed within 5 minutes at 56°C., leaving only a thermostable component, not affected by further heating. ( b ) The inactivation of K by this thermolabile component was not a first order reaction, but varied with the concentration of both serum and penicillin. At high concentrations of K, the rate of inactivation due to the thermolabile factor was negligible, and penicillin K was destroyed no more rapidly than F, G, or X. The rate of inactivation increased as the concentration of penicillin was reduced. At penicillin K concentrations of 50, 10, 2, and 0.4 micrograms per cc., the hourly destruction in rabbit serum averaged 10, 16, 21, and 54 per cent. The corresponding figures in human serum were 10, 11, 14, and 54 per cent. The reservations entailed by the large serum error at the lower concentrations of penicillin are discussed in the text. 4. The temperature coefficient for the inactivation of penicillin K by fresh human or rabbit serum was 2.5 for each 10°C. No significant inactivation was observed in 24 hours at 20°C.; and this was true also of penicillins F, G, and X. 5. Heparinized plasma was just as active as serum, washed red blood cells had no effect, and the activity of whole blood was referable to its plasma content. 6. The nature of the serum factors responsible for these two types of penicillin inactivation are under present study. 7. The urinary excretion of penicillin is so rapid that the slow destruction of penicillins F, G, and X in the circulating blood as here described is of secondary significance therapeutically. It nevertheless must contribute to their rapid disappearance from the blood; and the fact that X is inactivated more slowly than either F or G could be reflected in higher and more sustained blood levels than are afforded by the latter two species. There are some reports that such is the case (15–17), and the following paper provides further evidence for the superiority of penicillin X in this respect over the other species so far studied. The serum inactivation of penicillin K, at a rate which increases as its concentration falls, should be reflected in significantly lower and more evanescent blood levels than are observed with penicillins F, G, or X. As will be discussed in the following paper, this has been found to be the case, and provides a simple explanation for its paradoxically low therapeutic activity in vivo (8–11).

1. In both man and rabbit, penicillin X provided higher and more sustained blood levels than did penicillins F or G similarly administered in equal dosage (0.6 mg. per kg.); while penicillin K gave lower and evanescent levels. ( a ) One hour after intramuscular injection, the blood levels in rabbits averaged 0.11, 0.18, 0.02, and 0.33 for F, G, K, and X, respectively; and levels of 0.1 mg. per kg. were sustained for 1, 1.4,0.5, and 2.1 hours, respectively. ( b ) In man, the blood levels of G, K, and X averaged 0.23, 0.16, and 0.37 mg. per kg. 1 hour after intramuscular injections at 0.6 mg. per kg., and 0.11, 0.02, and 0.15 mg. per kg. 1 hour after intravenous injection. ( c ) In man, a level of 0.1 microgram per cc. was sustained for 1.6, 1.2, and 2.3 hours after the intramuscular injection of G, K, and X, respectively, and for 1, 0.5, and 1.3 hours after their intravenous injection. 2. The total urinary recovery of penicillins F, G, and X varied between 68 and 100 per cent, averaging 61, 87, and 74 per cent, respectively, in rabbits. In man, the urinary recovery of G and X averaged 86 and 93 per cent, respectively. In sharp contrast, the urinary recovery of penicillin K averaged 33 per cent in seven rabbits and 28 per cent in six human volunteers. The major portion of the penicillin appeared in the first 30 to 60 minutes. This suggests a rapid inactivation of penicillin K in vivo . 3. The therapeutic significance of these data is discussed in the text, and in greater detail in the following paper.

1. The relative bactericidal activities of penicillins F, G, K, and X against Type I pneumococcus in vitro were 60, 100, 180, and 135. The corresponding activities against Streptococcus pyogenes , strain C-203, were 75, 100, 115, and 145, respectively. 2. The total curative doses (CD 50 ) of penicillins F, G, K, and X in pneumococcal infections of white mice (ten injections at 3 hour intervals) were 4.6, 3.8, 20, and 2.4 mg. per kg., respectively, or relative activities of 83, 100, 19, and 160, referred to G as 100. 3. The corresponding curative doses in streptococcal infections of white mice were 2.6, 1.3, 14.0, and 0.5 mg. per kg., or relative activities of 50, 100, 9, and 260. 4. Penicillin K was therefore one-tenth as active in vivo as would be implied by its bactericidal activity in vitro . This probably reflects its rapid inactivation in vivo , evidenced by the low and evanescent blood levels observed in both rabbits and man, and the low urinary recovery of this species of penicillin. 5. Penicillin X was significantly more active therapeutically than its bactericidal activity in vitro would imply. This probably reflects its slower inactivation in vivo , evidenced by the somewhat higher and more prolonged blood levels afforded by this penicillin in comparison with penicillin G. Judged by the mouse infections with the strains here used, penicillin X is the penicillin of choice in the treatment of infections with pneumococcus Type I and hemolytic streptococci. 6. The curative dose of penicillin in streptococcal and pneumococcal infections paralleled the varying susceptibility of these organisms to penicillin in vitro .

1. Penicillin was found to be actively spirocheticidal in vitro against the Reiter, Kazan, Nichols, and Noguchi strains of so called S. pallida , and a strain of mouth spirochetes. The threshold concentration was 0.01 unit per Cc. (1–160,000,000 penicillin). The rate and degree of action increased with the concentration of penicillin up to a level of approximately 0.1 to 0.25 unit per cc., which rendered more than 99 per cent of the organisms non-viable within 12 hours. Higher concentrations did not appreciably accelerate the effect. 2. Within the range 4 x 10 4 –10 7 organisms per cc., the initial rate at which the spirochetes were killed was not affected by their number. Consistent with that observation, no demonstrable penicillin was bound or inactivated by thick suspensions. The amount of penicillin required to sterilize suspensions of varying density nevertheless varied to a large extent with the initial number of organisms. This was only in part due to the progressive deterioration of the penicillin with prolonged incubation; and the persistence of organisms resistant to the drug, and perhaps an adaptative change after prolonged exposure to penicillin, may be contributing factors. 3. The organisms remained actively motile for a period of 8 to 24 hours after they had been rendered non-viable by the action of penicillin. Even 500 units of penicillin per cc., or approximately 10,000 times an effectively spirocheticidal concentration, did not accelerate that delayed immobilization. It follows that, although penicillin rapidly renders the organisms non-viable, the metabolic system affected is not immediately essential to the life of the cell, and the motility and presumably other vital functions remain unaffected for a significant number of hours. 4. The rate at which the organisms were killed by penicillin increased with temperature in the range 8–40°C. With an original inoculum of 10 6 spirochetes per cc., the percentage of organisms surviving after 24 hours at 39–40°, 36–37°, 32–33°, 22–23°, and 8°C. was 0.02, 0.2, 1, 10, and 100 respectively; and those results were independent of the concentration of penicillin in the range 0.25 to 250 units per cc. If these observations with a non-pathogenic organism in vitro are applicable to the pathogenic organism in vivo , they suggest that the combined use of fever and penicillin in the treatment of syphilis may be more effective than either alone.

1. In confirmation of Gaehtgens, syphilitic human sera give positive complement fixation with cultures of so called T. pallidum (Reiter strain). Syphilitic rabbit sera are equally reactive. Syphilitic human and rabbit sera agglutinate these cultures, often in high titre (Beck). 2. Normal rabbit sera react weakly with the culture to give both agglutination and complement fixation in low titre. Normal human sera, despite the fact that they contain agglutinins in low titre, fail to fix complement with the Reiter strain of cultured spirochetes. Confirming Gaehtgens, the latter reaction is therefore of practical utility for the serum diagnosis of syphilis. 3. When syphilitic serum is heated at 63°C., there is no demonstrable difference in the thermolability of the antibody to spirochetes, and of the reagin which determines the Wassermann and flocculation tests. 4. ( a ) The absorption of syphilitic serum by spirochetal suspensions removes all reactivity, not only for the spirochetes, but for tissue lipoids (alcoholic beef heart extract) as well; the sera become Wassermann- and flocculation-negative. ( b ) Absorption of syphilitic serum with tissue lipoids renders the Wassermann and flocculation tests negative, but does not demonstrably change the reactivity of the serum with spirochetes. ( c ) Rabbits immunized to beef heart lipoid develop spirochetal agglutinins and complement-fixing antibodies (Reiter strain) in high titre. 5. It is concluded that these cultured spirochetes contain antigenic material serologically related to a substance present in mammalian tissue, as well as other antigenic factors not present in such extracts, but equally reactive with syphilitic serum. 6. These findings support the thesis that the primary serologic change in syphilis is the development of antibodies to T. pallidum . The Wassermann and flocculation tests would be explained on the basis that the tissue extracts used as "antigen" in these tests contain one or more substances serologically related to antigenic components of T. pallidum . Similarly, the cultured Reiter strain of spirochete is apparently sufficiently close serologically to T. pallidum to be agglutinated by and to give complement fixation with the antibodies to T. pallidum present in syphilitic serum. 7. Since suspensions of cultured spirochetes contain antigenic factors which react specifically with syphilitic serum, some of which are not present in ordinary Wassermann and flocculation "antigens," they may prove even more valuable than those tissue extracts in the serodiagnosis of syphilis.

Small amounts of formaldehyde inhibited the precipitating activity of horse diphtheria antitoxin with toxin and of horse antipneumococcus serum with the homologous capsular carbohydrate. Approximately 1 part of commercial formaldehyde to 1000 parts of serum, acting for 24 hours, inhibited the flocculating activity completely. In both cases, the combining affinity of the treated antibody for the corresponding antigen was not demonstrably affected, as determined both by in vitro experiments and by animal protection. More intensive treatment of the antipneumococcus serum caused an apparent loss of its bacterial agglutinating activity, but on centrifugation the organisms cohered: combination had occurred, and only the spontaneous aggregation was prevented. These effects are the same as those previously described for diazo compounds, and have been qualitatively reproduced with acetaldehyde, benzaldehyde and butyraldehyde. The quantitative relationships suggest that only a few groups in the antibody molecule need be modified by formaldehyde in order to prevent aggregation; and it is probable that these are some of the free NH 2 groups of the antibody protein. In marked contrast, the combining affinity of both antipneumococcus antibody and diphtheria antitoxin for the corresponding antigens was only slightly affected by amounts of formaldehyde which sufficed to block the free NH 2 groups rapidly and almost completely. Similarly, this amount of treatment did not affect the reactivity of these two antisera acting as antigen with a rabbit antiserum versus horse serum. The integrity of the NH 2 groups is apparently not essential for the activity of these sera acting either as antigen or as antibody; and the slow disappearance of their activity in concentrated HCOH is apparently to be ascribed to some secondary reaction other than the simple addition of HCOH to free NH 2 groups. The present experiments do not support the theory that antigen-antibody aggregates are lattice-like structures built up from elementary antigen-antibody compounds because of residual specific combining groups. The aggregating activity of both antipneumococcus serum and diphtheria antitoxin was completely inhibited by procedures which did not demonstrably affect their combining power with antigen. This suggests that the aggregation of antigen-antibody compounds is a secondary, non-specific reaction. It is perhaps significant that the amount of formaldehyde which just sufficed to prevent aggregation also caused a marked increase in the solubility of the pneumococcus antibody, which could then no longer be precipitated at serum pH by dilution with water or by dialysis. This strongly suggests that the loss of precipitating activity is actually due to the increased solubility of the antibody and supports the hypothesis that the primary cause of specific antigen-antibody aggregation is the relative insolubility of the bound antibody.