1. The process of fusion between two red corpuscles so as to form a round or slightly oval macrocyte with normal hemoglobin content, a diameter of 10µ± and no delle.

These macrocytes appear to be slightly thicker in the centre than at the periphery. No thickness measurements could be made.

2. The process of fusion between two microcytes each 5µ± in diameter, so as to form a red corpuscle of approximately normal size but with dense hemoglobin and no delle.

3. The process of fusion between a red corpuscle and a microcyte, forming a macrocyte slightly larger than a normal red corpuscle and exhibiting a normal hemoglobin tint but no delle.

4. The process of microcyte formation from red corpuscles was seen occurring in four different ways: (a) By the intermittent compression of a single red corpuscle by the pull of three or more fibrin threads attached to the corpuscle. (b) By the steady, continuous compression of a single red corpuscle by the pull of two polar fibrin threads, (c) By intermittent rises of intracorpuscular pressure in a red corpuscle due to trauma, (d) By avulsion and retraction or contraction of the hemoglobin stroma from the intact surface layer of a red corpuscle.

Thickness measurements in single microcytes were made in rabbit blood.

5. The process of a typical microcyte with dense hemoglobin and no delle changing back to a fairly large red corpuscle with practically normal hemoglobin tint and with a delle.

6. Macrocytes were seen expelling their hemoglobin in a fraction of a second and turning into shadows which were often paler than the background, but without any decrease in the visible surface area or change in the contour.

7. Extensibility and retractibility (or perhaps relaxation and contraction) may be exhibited by the corpuscle as a whole, by local sections of both surface layer and stroma, or retractility (contractility (?)) may be shown by the hemoglobin stroma alone.

On the basis of observational evidence it is inferred that:

1. The surface layer of a red corpuscle is probably a condensed stroma rather than an anatomically defined membrane.

2. A pigment, probably hemoglobin, is present in the surface layer.

3. The surface layer of red corpuscles becomes rigid before or during spontaneous hemolysis.

4. Hemoglobin exists in an undissolved form in the intact red corpuscle.

5. The transverse diameter of microcytes (5µ±) represents the maximal shortening of which this diameter is capable in normal red corpuscles of man, rabbit and guinea pig under the conditions studied.

6. Fusing red corpuscle are destroyed by repeated washings with Ringer solution.

7. In stagnating blood there is no increase in fusion forms, nor in the production of microcytes.