It is usual to take the blood from the tip of the finger, and only in exceptional cases, _e.g._ in oedema of the finger, are other places chosen, such as the lobule of the ear, or (in the case of children) the big toe. For the puncture pointed needles or specially constructed instruments, open or shielded lancets, are unnecessary: we recommend a fine steel pen, of which one nib has been broken off. It is easily disinfected by heating to redness, and produces not a puncture but what is more useful, a cut, from which blood freely flows without any great pressure. The literature dealing with the numbers of the red corpuscles in health, is so large as to be quite unsurveyable. According to the new and complete compilation of Reinert and v. Limbeck, the following figures (calculated roundly for mm.^{3}) may be taken as physiological: _Men._ Maximum Minimum Average 7,000,000 4,000,000 5,000,000 _Women._ Maximum Minimum Average 5,250,000 4,500,000 4,500,000 This difference between the sexes first makes its appearance at the time of puberty of the female. Up to the commencement of menstruation the number of corpuscles in the female is in fact slightly higher than in the male (Stierlin). Apart from this, the time of life seems to cause a difference in the number of red corpuscles only in so far that in the newly-born, polycythaemia (up to 8-1/2 millions during the first days of life) is observed (E. Schiff). After the first occasion on which food is taken a decrease can be observed, and gradually (though by stages) the normal figure is reached in from 10-14 days. On the other hand the oligocythaemia here and there observed in old age, according to Schmaltz, is not constant, and therefore cannot be regarded as a peculiarity of senility, but must be caused by subsidiary processes of various kinds which come into play at this stage of life. The influence which the taking of food exercises on the number of the red blood corpuscles is to be ascribed to the taking in of water, and is so insignificant, that the variations, in part at least, fall within the errors of the methods of enumeration. Other physiological factors: =menstruation= (that is, the single occurrence), =pregnancy=, =lactation=, do not alter the number of blood corpuscles to any appreciable extent. The numbers do not differ in arterial and venous blood.