which enters into relation with the maternal tissue, sends out numerous processes which permeate the epithelial layer and the stroma of the mucous membrane in its immediate vicinity, causing liquefaction and destruction of the maternal tissue, so that very soon the ovum is almost entirely surrounded by mucous membrane of the uterus. The trophoblast in places becomes vacuolated forming a system of intercommunicating lacunæ which become bathed in maternal blood by the erosive action of the trophoblast against the walls of the dilated capillaries, which later form the placental sinuses. Into the trophoblastic processes there soon grows a core of mesoblastic tissue bearing fetal capillaries from the umbilical blood-vessels and the cells of the trophoblasts arrange themselves more or less irregularly into two layers-the inner or Langerhans' layer, and an outer or plasmodial layer generally called the syncitium, thus forming the true chorionic villi.

To combat and to limit the activity of the embedded ovum, the connective-tissue cells surrounding the blastocyst proliferate to form decidual cells and the superficial capillaries of the mucous membrane dilate enormously. The epithelium at the end of the chorionic villi in relation to maternal tissue is thickened to form "root caps" or "Haftzotten." As the ovum grows, that portion of the decidua which hems in the blastocyst and is generally termed the decidua reflexa" becomes stretched and thinned, the blood supply becomes very limited and the chorionic villi in relation to it cease to grow and later degenerate and atropyh; whereas the villi in the region of the basal decidua―or as it is commonly called the decidua serotina-continue to grow and enlarge and become branched and thus the discoid placenta is formed. Occasionally the chorionic villi in relation to the decidua reflexa near the serotinal border continue to grow forming the so-called reflexal placenta. Later in pregnancy many of the villi become fibrinized because of obliterative changes in the vessels of the villi.

While the placenta is in process of formation, changes in the mucous membrane of the rest of the uterus occur which are similar though less marked-to those in the serotina, i.e., degeneration of surface and glandular epithelium, dilatation of capillaries and hyperplasia of the connective-tissue stroma to form the decidua vera.

With this idea of the development of the placenta in its normal position, i.e., on the anterior or posterior surface and near the fun

dus of the uterus, we proceed to the consideration of its abnormal implantation upon the peritoneum, in the ovary, in the Fallopian tube and broad ligament, in a malformation of the uterus and in the wall and within the cavity of the uterus.

Upon the Peritoneum.-When we consider the active resistant and absorbing power of the peritoneum, it seems almost impossible for an ovum to become primarily implanted upon this structure; and experiments upon animals of ova transplanted into the peritoneum would seem to bear out this hypothesis. However, Galabin in 1896, before the London Obstetrical Society, reported a case which a committee decided was probably "an example of primary abdominal gestation." In 1903, Wittbauer published a similar case and in 1908 B. C. Hirst reported a specimen which he maintains answered the following requirements of primary implantation, viz.:

1. Normal condition of tubes, ovaries and broad ligament, except where ovum is implanted.

2. No penetrations of intraligamentous space from ovarian fimbria.

3. No intraligamentous rupture of tube.

4. Proof that peritoneum constitutes reflexa of ovum. Inasmuch as no specimen so far reported has attained an age sufficient for the formation of a distinct placenta, we shall not discuss the question in this paper.

In the Ovary. The cases of ovarian implantation of the ovum seem to show the following factors: Fertilization, retention and growth of ovum in Graafian follicles or in the immediate vicinity of the follicle, active proliferation of the trophoblast and burrowing of the ovum within the stroma of the ovary with marked destruction of ovarian tissue and hemorrhage and with very little or no reaction of the connective-tissue cells of ovary to form a decidua; the destructive activity of ovum is so marked that the ovarian capsule is soon eroded, hemorrhage occurs which either causes death of patient or necessitates operative removal of ovary involved; there is no tendency to true placental formation.

The cases of ovarian pregnancy which have been described as advancing to term with the formation of a placenta have probably not been primary ovarian implantations, as all positive ovarian pregnancies so far described have shown early destruction and rupture; from our present knowledge of placental formation it is difficult to conceive of the development of a true placenta in a tissue incapable of decidual reaction.

Whether the embedding of the fecundated ovum in the Fallopian tube is due to inflammatory changes within or external to the tube wall itself or in the ciliated epithelium, or to newgrowths, or to diverticula in the tube, or to developmental anomalies, or to psychical conditions; or whether it is due to changes in the ovum itself, is not in the province of this paper to discuss; we shall consider only distinct placentation in relation to the Fallopian tube. The ovum of Peters, those of v. Spee, the collected ova described by Webster and the recent work of Bryce and Teacher, taken in connection with specimens of tubal pregnancy, seem to prove that distinct placental formation can occur only in those tissues capable of reacting to form a decidua, thus limiting the activity of the trophoblast of the chorionic villi; and that, if this decidual formation is slight or absent, the ovum must be a sacrifice to its own activity, caused by destruction of maternal tissue and hemorrhage.

While the mucous membrane of the tube is genetically similar to that of the uterus, the tubal reaction to the implanted ovum is not at all constant; in some the decidual formation is very scant or absent, while in other specimens it is marked and sufficient for the development of a typical decidua serotina. When the decidual formation is slight, a membrana capsularis develops which, however, is soon eroded, leading to rupture or abortion; when the decidual reaction is marked, a true placenta is formed and, as the fetus develops, the structures of the tube become so thinned that it soon ruptures either into the peritoneal cavity, forming a tuboabdominal pregnancy, or into the broad ligament, forming a tubointraligamentous pregnancy. These latter conditions may be associated with so much hemorrhage that either the fetus or mother or both may die; or, if the placenta has not been disturbed to any extent, the fetus in its sac may continue to grow to term and may by operative interference be born alive. More often the fetus dies in the later months and may undergo calcareous, fatty, or putrefactive changes; the intimate association with the intestines makes infection by the colon bacillus especially to be feared.

When the ovum develops in that portion of the tube within the uterine wall, there is always the possibility of the fetus being dislodged into the uterine cavity, the placenta remaining within the wall. Because of the thickness of the uterine wall these interstitial ectopics often reach a more advanced stage before rupturing than those in any other portion of the tube.

The diagnosis of the various ectopics with placental formation is often a difficult matter and they have often enough been mistaken for intrauterine pregnancies. The question is too involved to be considered in this paper.

In regard to the treatment of these cases, the general law in reference to all extrauterine pregnancies holds good; i.e., when the diagnosis is made, operate. It was formerly held to be the better plan to wait with the hope that the patient would have a spurious labor, that the fetus would die and that, after waiting several days, an operation could remove both fetus and placenta without the large amount of hemorrhage associated with the removal of a living placenta. While the difficulties of the operation are lessened by this procedure, the dangers during the waiting period, e.g., spontaneous hemorrhage, infection, outweigh the advantages, and the preponderance of opinion to-day favors early operation. Whether the sac be removed with the fetus and whether the placenta be removed at the primary operation or be deferred for several days, are matters to be decided by the conditions and anatomical relations found in each particular case. In general, the abdominal route is preferred to the vaginal by most operators.

That pregnancy may occur in the rudimentary horn of a bicornuate uterus has been proved by cases collected by Sänger, Himmelfarb and Kehrer; they generally proceed to placental dedevelopment and may go to term; the tendency to rupture and hemorrhage is, however, great, especially if the horns of the uterus are completely separated; the indication is for an operation as soon as diagnosis is made.

In the Uterus.-The normal site in the uterus for the implantation of the placenta may be said to be anywhere in the upper uterine segment, the favorite location being the posterior or anterior walls and extending upon the fundus; there is some difference of opinion as to which position is the more common, but there is probably no great variation, as Holzapfel in a series of 107 cases found the placenta on the anterior wall in forty-two and on the posterior in forty-five. The implantation in the horn of the uterus is not uncommon, and in the quoted series the placenta was found fourteen times in the tube angle and five times laterally just beneath a tube.

An implantation within the lower uterine segment may be said to be abnormal, although many of these situated at some distance from the internal os give rise to no symptoms and have

been termed "vicious implantation" by Pinard who, judging from the location of the rupture in the membranes, stated that they occurred in over 25 per cent. of all cases; this large percentage has not been generally accepted as based on fact.

When the plancenta is situated over the internal os or when during the dilatation of the cervix any portion of the placenta overlaps the circumference of the cervix, it is termed a placenta. previa, a condition not uncommonly found, although there is no unanimity as to the frequency, the reports varying from one case in 133 delivered to one in 1500, the greater frequency occurring in hospital practice. It is generally admitted that the condition is seldom found in primiparæ, Galabin stating that only 4.4 per cent. of the London and Dublin cases were primiparæ. Of the different forms of placenta previa it is generally admitted that the partial variety is of much greater frequency than either the central or marginal.

When we consider the cause of placenta previa, we enter a field of speculation. Hofmeyer's theory, that the condition is due to the continued growth of the chorionic villi in relation to part of the decidua reflexa, forming a reflexal placenta, and the later union with the decidua vera in the region of the internal os, is accepted as a cause in some cases. Many believe that a placenta previa can occur only from a primary implantation in the lower segment of the uterus and its extention over the internal os by cleavage of decidua vera. That the ovum may be implanted at the internal os has been denied by eminent authorities, on the ground that no specimen has ever been shown in which the ovum was implanted in this locality and that it was impossible because of the small size of the ovum as compared with the internal os. Bumm, however, in answer to this asserts that the walls of the cavum uteri, as well as the internal os, are in apposition forming a capillary tube and that the ovum is prevented from adhering to the upper segment by muscular action of the uterus or by the lack of adhesive power of a chronically inflamed mucous membrane and that its expulsion from the uterus is retarded by a plug of mucus and blood at the internal os, and that the burrowing ovum in this location soon causes degeneration of the epithelial covering at this site and the coalescence of the walls and the formation of the decidua serotina; and he also shows that findings in mammals with placentation similar to the human, prove the possibility of a placenta at the internal os formed in this manner.