The configuration assumed by the outflowing blood cells depends upon the consistency of the medium and the elevation of the fragment above the general level of that medium. If the medium is composed of serum, the cells flow out from all sides as when a vessel bursts that contains a viscous fluid. In the presence of plasma, however, the outflowing cells may take every conceivable form ranging from broad, fan-like disseminations, through short, stalky, bud-like projections to long, slender ones. It will not be possible at this time to describe more thanone type of forma-tion, namely, that which results in the subsequent development of a long, slender tube. At one or more points on the margin, a few cells become dislodged and break away from the fragment. Their places are taken by those behind. If the proximal impact is great, these cells are also pushed out into the medium. A general streaming begins. Each cell that is forced out follows in the wake of those that have gone before. They may proceed in single file, or abreast of one another, but always over exactly the same route. This route assumes the nature of a tunnel-like passage through the plasma. As long as the cells in the lead are being pushed forward by those from behind, they will continue to advance through the medium until it coagulates. The moment coagulation occurs, however, the cells in the lead, being unable to make further progress, come to a stop. If coagulation occurs before the proximal outflow has ceased, the passage will remain open and unobstructed, despite the fact that the cells in transit may be widely separated If the pathway is only partially filled with cells, they may be made, at any time, to flow in either direction by manipulating the flask. After this happens, the force of the outflowing cells may still be so great that a large spherical expansion is formed at its distal end. Enlargements may also appear elsewhere along the tube. Very often it may be completely ruptured, and occasionally, short branches are found.

All this may occur within half an hour after the cultures have been prepared. As soon as the sur-rounding medium has become firmly coagulated, however, now further change occurs either in the length of the tube or in its diameter. The structures do not, as Hueper and Russell2 have reported, continue to grow upon further incubation. Up to this point, the development of the tubules depends solely upon the rapidity with which the plasma coagulates. A culture held at room tempera-