And biochemists make an effort to ‘keep up with the literature’. Every paper lists a dozen or so other papers in the field. Every review article lists a hundred to a thousand. For instance, the article on serotonin which I mentioned earlier in the paper includes a listing of one hundred and fifty-seven papers, or ‘references’ as they are called. The one on microheterogeneity of proteins includes one hundred and sixty-four references.

  But where do writers of papers and, particularly, of review articles get their lists of references? Fortunately, there are journals which devote themselves to nothing more than preparing ‘abstracts’ of scientific papers. That is, with the aid of an army of technically trained people willing to work for the good of science and a nominal sum, the journal will try to keep track of every paper appearing in every journal related to their field. They will list for every paper all over the world, the title, author or authors, name, volume, month or page of the journal, and a short summary of the paper’s contents. The most important such journal for our purposes is Chemical Abstracts, which abstracts all foreign papers in the very best English.

  Chemical Abstracts comes out twice a month. An individual issue has up to four hundred large pages in double columns and microscopic print. Each column is numbered separately beginning with the first issue of a year and ending with the last. The total number of columns per year—of the listing of papers only—used to reach eight thousand ten years ago. It now reaches seventeen thousand.

  Every year Chemical Abstracts publishes an extensive and exhaustive author index and subject index. They come out in three volumes and add up to more than a thousand pages. When I first became interested in such things, the index came out three months after the year ended. As it has grown larger, longer, and more complicated, it comes out now nearly a year after the year ended. This means that there is always a minimum of say twenty issues and a maximum of forty issues of Chemical Abstracts unindexed. These unindexed issues are the latest ones which contain the latest papers.

  This means that if you’re trying to read up on the work done in a given field, you first exhaust the indices for the last ten years, say, then get grimly to work on the individual issues of the last year or two.

  And when you’ve got the entire field of biochemistry to keep up with for the sake of a textbook—ouch, each aching vertebra.

  Review articles are a boon and a gift from the gods, but even one which is freshly printed, and which contains the latest information, can’t include all the papers in the field or do more than refer very cursorily to most of those it does include. It never hurts to do a little browsing through Chemical Abstracts on your own, which, by the way, lists all review articles. Furthermore, the number of review articles being published now is so great, that you can’t read all you should of those either.

  In preparing future editions of the text, the one great problem is ‘bringing it up to date’ and for that Chemical Abstracts is absolutely necessary. My own method is to grab Dr. Walker’s issue of Chemical Abstracts (he subscribes!) as soon as it comes in, preferably before Dr. Walker gets his hooks on them. Fortunately, Chemical Abstracts segregates its paper listings into over twenty subdivisions of chemistry and I can ignore sections dealing with industrial chemicals, paper and paper-making chemicals, electrochemistry, photography and so on. Unfortunately, the listings under Biochemistry—itself subdivided into ten subdivisions—is the longest in the periodical.

  I cuddle up with one hundred to one hundred and fifty large pages containing one thousand to one thousand five hundred articles twice a month, in other words, and read dizzily through the titles. Sometimes a title is short, like ‘Iron Metabolism’ which usually indicates a review. (All reviews are automatically noted down by me provided they are in a journal I can obtain. In one place or another in Boston I can obtain almost all the unimportant journals and all the important ones. I can obtain almost all the important ones by going to the school library two floors below my lab.)

  Sometimes the title is long, like for instance: Use of Ion Exchangers for the Separation of some of the Amino Acids formed during the Enzymic Degradation of Cysteinesulfinic Acid. Application to the Isolation of Hypotaurine (2-Aminoethanesulfinic Acid), which is the real title of a real paper. Long titles like this are fashionable not because scientists are queer, but because a good title is one which will give you a complete idea of the contents of the paper, without your having to read anything further. That’s not laziness on our part, friend, that’s one of our barriers against insanity.

  If a title of a paper is interesting, I read the abstract itself. If the abstract looks interesting, I note the volume of Chemical Abstracts, the number of its column and its position in that column in a special volume of our textbook with a blank page between every two printed pages. I make the entry opposite the place in the book where I think it belongs.

  The results? Well, they can be harrowing. Take the case of the function of the metal, molybdenum, in the human body. In the first edition of our book, it wasn’t even worth mentioning and we didn’t mention it. By the time we wrote the second edition, some workers had showed it to be a constituent of an important enzyme known as xanthine dehydrogenase. We stuck in molybdenum, therefore, and gave it seven lines. By the time the third edition rolled round there were thirty new papers to be read, or at the very least, glanced through, in order that we might increase the space devoted to molybdenum from seven lines to two paragraphs, and do it intelligently. And this despite pruning the number ruthlessly by first picking only those with interesting titles and, of those, only the ones with interesting-sounding abstracts.

  And this isn’t really enough, you know. Even Chemical Abstracts isn’t up to date. They’re anywhere from six months to a year behind the journals. One ought, therefore, to glance at the titles in the most important journals as they come out. But then, the journals aren’t up to date, either. A paper which is accepted for publication by the Journal of Biological Chemistry may have to wait six months to a year before seeing the light. The journal has that great a backlog of accepted papers. Besides that, a paper deals with work that is completed. There is other, newer work in progress.

  And so there are all sorts of conventions. The American Chemical Society holds annual conventions in various parts of the country. The Federation of American Societies for Experimental Biology—which includes six subsidiary societies—holds annual conventions. So does the American Association for the Advancement of Science. So do innumerable smaller groups. At each one of these, papers are presented. Hundreds of papers are presented at the largest gatherings, where several subgroups are usually giving series of papers simultaneously in different rooms of the hotel—sometimes in different hotels and sometimes even in different cities. It is impossible for one man to hear more than a fraction of these and he must choose his spots with care and hope for the best.

  Of the three of us, Dr. Boyd is the most indefatigable attender of conventions. In recent months, he has been to New York, Philadelphia, French Lick (Indiana) and Paris (France) in order to give papers, listen to papers and—most important of all—get together over a glass of beer and find out what’s doing in the other guy’s lab right at that moment.

  And so it goes.

  There is now a whole branch of human effort devoted to attempts to coordinate the accumulating data of the physical sciences at a rate roughly equivalent to that at which it is accumulating. This includes the formulation of special types of indices and codes, the use of screening programs, the preparation of special punched cards, micro-card files and so on.

  In connection with this, I should like to quote a passage from a letter written by Karl F. Heumann, Director of the Chemical-Biological Coordination Center of the National Research Council to Mr. Ken Deveney, Jr., of Millington, New Jersey. A carbon copy was sent to John Campbell, the editor of Astounding Science Fiction, who forwarded it to me.

  The passage reads:

  Dear Dr. Deveney:

  In answer to your question...about data-handling. I would like to give you a short bibliography but it is not possible. There has been a great increase in work in this field which has resulted in a scattering of documentation references among the various abstracting services....

  In other words, the literature relating to efforts to handle the literature is too great to be handled easily.

  Get it?

  There’s no way out and each year it’s getting worse.

  —And so, if you are ever up Boston way, and enter the Boston University School of Medicine and pass my lab and hear the sound of panting, you may think it is the result of my chasing some female around and around some desk—but you’d be wrong.

  It’s just Asimov trying to keep up with the literature, a task which is much more futile and far less likely to reach a satisfactory conclusion.

  This article was first written in December 1954, when the second edition of our textbook had just appeared. A third edition, in which the revisions were carried through mostly by myself, appeared in 1957. In that year, however, Dr. Walker retired, and the next year I, myself, retired from active teaching. (Drs. Walker and Boyd are both in their seventies at this moment of writing, but both are still alive and vigorous.) With all this, the textbook died. I am sorry to say.

  As for the matter of the scientific literature, I assure you that everything described in the article has grown still more unwieldy. The scientific literature is supposed to double its rate of production every ten years, which means things are four times as bad now as they were when I wrote this article. I think I’m glad I’m not working on the textbook anymore.

  SPECIAL NOTE

  The two chapters that follow, unlike the preceding, are not legitimate reporting and speculation. Rather they are a gentle spoofing of science and scientific papers.

  In the case of each, one outrageous assumption is made, so that we are presented with a most unusual chemical compound in the first and an equally unusual bird in the second. Once this assumption is made, everything else follows more or less plausibly. Don’t let this plausibility fool you, however, into taking either article seriously.

  CHAPTER ELEVEN—THE MARVELLOUS PROPERTIES OF THIOTIMOLINE

  PART I

  The correlation of the structure of organic molecules with their various properties, physical and chemical, has in recent years afforded much insight into the mechanism of organic reactions, notably in the theories of resonance and mesomerism. The solubilities of organic compounds in various solvents has become of particular interest in this connection through the recent discovery of the endochronic nature of thiotimoline.{2}

  It has been long known that the solubility of organic compounds in polar solvents such as water is enhanced by the presence upon the hydrocarbon nucleus of hydrophilic—i.e. water-loving—groups, such as the hydroxy (-OH), amino (-NH2), or sulfonic acid (-SO3H) groups. Where the physical characteristics of two given compounds—particularly the degree of subdivision of the material—are equal, then the time of solution—expressed in seconds per gram of material per milliliter of solvent—decreases with the number of hydrophilic groups present. Catechol, for instance, with two hydroxy groups on the benzene nucleus dissolves considerably more quickly than does phenol with only one hydroxy group on the nucleus. Feinschreiber and Hravlek{3} in their studies on the problem have contended that with increasing hydrophilism, the time of solution approaches zero. That this analysis is not entirely correct was shown when it was discovered that the compound thiotimoline will dissolve in water—in the proportions of 1 gm./ml.—in minus 1.12 seconds. That is, it will dissolve before the water is added.

  Previous communications from these laboratories indicated thiotimoline to contain at least fourteen hydroxy groups, two amino groups and one sulfonic acid group.{4} The presence of a nitro group (-NO2) in addition has not yet been confirmed and no evidence as yet exists as to the nature of the hydrocarbon nucleus, though an at least partly aromatic structure seems certain.

  The Endochronometer—First attempts to measure the time of solution of thiotimoline quantitatively met with considerable difficulty because of the very negative nature of the value. The fact that the chemical dissolved prior to the addition of the water made the attempt natural to withdraw the water after solution and before addition. This, fortunately for the law of Conservation of Mass-Energy, never succeeded since solution never took place unless the water was eventually added. The question is, of course, instantly raised as to how the thiotimoline can ‘know’ in advance whether the water will ultimately be added or not. Though this is not properly within our province as physical chemists, much recent material has been published within the last year upon the psychological and philosophical problems thereby posed.{5} {6}

  Nevertheless, the chemical difficulties involved rest in the fact that the time of solution varies enormously with the exact mental state of the experimenter. A period of even slight hesitation in adding the water reduces the negative time of solution, not infrequently wiping it out below the limits of detection. To avoid this, a mechanical device has been constructed, the essential design of which has already been reported in a previous communication.{7} This device, termed the endochronometer, consists of a cell 2 cubic centimeters in size into which a desired weight of thiotimoline is placed, making certain that a small hollow extension at the bottom of the solution cell—1 millimeter in internal diameter—is filled. To the cell is attached an automatic pressure micro-pipette containing a specific volume of the solvent concerned. Five seconds after the circuit is closed, this solvent is automatically delivered into the cell containing the thiotimoline. During the time of action, a ray of light is focused upon the small cell-extension described above, and at the instant of solution, the transmission of this light will no longer be impeded by the presence of solid thiotimoline. Both the instant of solution—at which time the transmission of light is recorded by a photoelectric device—and the instant of solvent addition can be determined with an accuracy of better than 0.01 per cent. If the first value is subtracted from the second, the time of solution (T) can be determined.

  The entire process is conducted in a thermostat maintained at 25.00° C.—to an accuracy of 0.01° C.

  Thiotimoline Purity—The extreme sensitivity of this method highlights the deviations resulting from trifling impurities present in thiotimoline. (Since no method of laboratory synthesis of the substance has been devised, it may be practically obtained only through tedious isolation from its natural source, the bark of the shrub Rosacea Karlsbadensis rufo.{8}) Great efforts were therefore made to purify the material through repeated recrystallizations from conductivity water—twice redistilled in an all-tin apparatus—and through final sublimations. A comparison of the solution times (T) at various stages of the purification process is shown in Table I.

  It is obvious from Table I that for truly quantitative significance, thiotimoline purified as described must be used. After the second resublimation, for instance, the error involved in an even dozen determinations is less than 0.7 per cent with the extreme values being -1.119 seconds and -1.126 seconds.

  In all experiments described subsequently in this study, thiotimoline so purified has been used.

  Time of Solution and Volume of Solvent—As would seem reasonable, experiments have shown that increasing the volume of solvent enables the thiotimoline to dissolve more quickly—i.e. with an increasingly negative time of solution. From Figure 1, however, we can see that this increase in endochronic properties levels off rapidly after a volume of solvent of approximately 1.25 ml. This interesting plateau effect has appeared with varying volume of solvent for all varieties of solvents used in these laboratories, just as in all cases the time of solution approaches zero with decreasing volume of solvent.

  Time of Solution and Concentration of a Given Ion—In Figure 2, the results are given of the effect of the time of solution (T) of varying the volume of solvent, where the solvent consists of varying concentrations of sodium chloride solution. It can be seen that, although in each case the volume at which this plateau is reached differs markedly with the concentration, the heights of the plateau are constant (i.e. -1.13). The volume at which it is reached, hereinafter termed the Plateau Volume (PV), decreases with decreasing concentration of sodium chloride, approaching the PV for water as the NaCl concentration approaches zero. It is, therefore, obvious that a sodium chloride solution of unknown concentration can be quite accurately characterized by the determination of its PV, where other salts are absent.

  This usefulness of PV extends to other ions as well. Figure 3 gives the endochronic curves for 0.001 molar solutions of sodium chloride, sodium bromide, and potassium chloride. Here, the PV in each case is equal within the limits of experimental error—since the concentrations in each case are equal—but the Plateau Heights (PH) are different.

  A tentative conclusion that might be reached from this experimental data is that the PH is characteristic of the nature of the ions present in solution whereas the PV is characteristic of the concentration of these ions. Table II gives the values of Plateau Height and Plateau Volume for a wide variety of salts in equal concentrations, when present alone.

  The most interesting variation to be noted in Table II is that of the PV with the valence type of the salt present. In the case of salts containing pairs of singly-charged ions—i.e. sodium chloride, potassium chloride, and sodium bromide—the PV is constant for all. This holds also for those salts containing one singly charged ion and one doubly charged ion—i.e. sodium sulphate, calcium chloride, and magnesium chloride—where the PV, though equal among the three, varies markedly from those of the first set. The PV is, therefore, apparently a function of the ionic strength of the solution.