This section is from the book "Materia Medica And Therapeutics Inorganic Substances", by Charles D. F. Phillips. Also available from Amazon: Materia medica and therapeutics.
According to Nasse, there exists in 1,000 parts of blood, 0.832 of iron oxide. Haemoglobin contains 0.42 per cent. as a constant quantity; most of it is in direct organic union with the red corpuscles, in the proportion of about 1 part of iron to every 230 (Gorup Bes-anez); when dried, they contain seven times as much as the fibrine, and four times as much as the serum (Boussingault). Being required then for the normal constitution of red blood, iron is essentially a food, but since illness follows deficiency in the number or quantity of corpuscles, and iron in substance will often remedy such illness, it equally comes within our province as a medicine, and from its curative effects, we may, inverting the general rule, deduce some part of its physiological action. That it can increase the number of red corpuscles is shown by the observations, e.g., of Rabuteau, who counted them by Malassez's method, in a case of chlorosis before and after twenty days' treatment by protochloride of iron: he found the number in a cubic millimetre to be nearly doubled (Gazette des Hopitaux, January, 1875); and in a specimen analyzed by Prof. Simon, the globulin and haematin were more than trebled ("Animal Chemistry," Sydenham Society). I need not multiply examples of this fact (though it has been denied), but there is something further to be learnt from the recent and careful observations of Hayem, on the blood of anaemic persons (Comptes Rendus, 1876, p. 985). He found that in cases of moderate chlorosis, the number of corpuscles was not markedly less than normal, but they were altered in shape and size, apparently in consistence, but most markedly in color-power, so that a given quantity showed a red tint not deeper than that of half the number of normal corpuscles. Further, after a course of iron, the number of corpuscles in the same patient was not always increased, sometimes it was diminished, but then the corpuscles individually had grown larger and of normal shape, and of so good a color as to equal even a greater number of the ordinary kind; he concludes then that iron acts by improving the internal nutrition of the globules, "it solicits them" to take up more haematin, more coloring matter. These observations confirm the older ones of Le Canu (These, 1847), that iron is the main constituent of haematin, is inseparable from the coloring matter, and must be at least an important element in the color itself. Hayem's conclusions are of still more importance as bearing on the assertions of Denis and of C. Bernard, that there is no real deficiency of iron in chlorotic blood, because they prove such a definite change in its vital characters under the medicinal use of the drug. Granted that there is no numerical, there is clearly a physical or a vital change produced by iron; and although it may be true that ordinary nutriment contains as much iron as should be wanted (Bernard), yet it seems equally true that we may sometimes have to give much that we may get a little absorbed (Gubler), that we must therefore give it "en masse," as we do, and (apart from all theory) Hayem furnishes us with a rational basis for our therapeusis. That the proportion of iron can vary in blood is proved by the analyses of Picard (Comptes Rendus, November, 1874); in 100 c.c. taken from three dogs respectively young, adult, and weakened by hemorrhage, he found that the amount of iron was .092, .065, and .041, and he established also the fact of a definite and constant relation between the amount of iron in any specimen of blood, and the amount of contained oxygen as liberated in vacuo from quantities of 100 c.c.
If it be asked how iron adds itself to the corpuscles and promotes their growth, we must recognize that it is not by mechanical addition to the formed corpuscle, or else the proportion in chlorotic blood could be at once increased, and failure to cure would not occur, nor relapse be so frequent. An observation by Quevenne throws some light upon the process; he found in proteid solutions withdrawn from the stomach of dogs more abundant precipitates of nutrient material if meat or wine, or iron especially had formed part of a meal, and suggested that in the portal vein a similar precipitate occurs (from the meeting of currents from splenic and mesenteric vessels both laden with the results of digestion in intimate contact with the added iron), that such precipitate is at least precursory to the formation of globules, and that at this stage iron exerts its blood-forming power ("Memoire," etc.). It would seem that better corpuscles are formed when (the vital processes being fairly active) one of their essential constituents is presented in unusual abundance for absorption. It becomes then combined with them in some organic, rather than chemical or mechanical union; and besides such direct action in the formation of globules, iron exerts special stimulating power over the blood-glands, which power, indeed, is by Trousseau and others considered more important than the last-mentioned. Further, when even a few new corpuscles have been formed, they add fresh nerve-energy and improve digestion, and the blood-forming process becomes still more actively assisted. Iron has been variously thought to be in the corpuscles in its metallic state, as phosphorus exists in the brain (Le Canu, Mulder), or as a free phosphate (Fourcroy), or as a peroxide (Denis, Mialhe, and a majority of observers). A precise chemical theory was elaborated by Liebig, who taught its presence as peroxide on account of its reactions with sulphuric acid, and found that this hydrated peroxide, in contact with moist organic membrane in partly closed vessels, could change to a protocarbonate, and on free exposure to oxygen could change back again, with evolution of carbonic acid: so that venous blood was held to contain a protocarbonate, and recently aerated blood a peroxide. It is difficult to accept so entirely chemical a theory, which implies that the element is more loosely combined with the corpuscle - more distinct from its substance - than it can be; other difficulties are stated in physiological works, and Liebig's view, though highly ingenious, and containing no doubt a partial truth, can only be accepted as an hypothesis.