[Introduction]
Up until a few years ago there was very little published in the medical literature on the physical findings consistent with child sex abuse. There was even less published on normal findings in non-abused children.

Moreover, there were no extensive studies on what the hymen of the non-abused child looks like. Some physicians claimed that the normal hymen is circular and that any irregularity meant something had been inserted into the vaginal canal. Although others maintained that there is a wide variety of irregularities within the normal range, they were unable to provide specific experimental data regarding the frequency of these irregularities.

There was even controversy regarding the size of the normal hymenal ring at various ages. Until recently, there were no extensive studies in which measurements were taken. And even the studies that were done were flawed by the fact that the investigators failed to consider that the hymenal orifice varies in size with the position the child assumes when the examination is being conducted, as well as with the degree to which the child's legs are spread by the examiner.

Similarly, there was no good information regarding the differences between the normal anus and the anus that has been subjected to sexual abuse. The necessary baseline studies had not been done.

In spite of this relative ignorance, physicians have been asked with increasing frequency to provide the definitive "proof" regarding whether or not sexual abuse has taken place. This has been the case even though most people who are knowledgeable about child sex abuse recognize that often there will be no physical findings because the perpetrator has not done anything more than caress and fondle the child. However, the need for such verification has been strong, so strong that the objectivity of both those who make the request and those physicians who have responded has been compromised.

In response to this need, pediatricians, pediatric gynecologists, and people from other branches of medicine (such as internal medicine and family practice) have become "experts" on child sex abuse in recent years. Those who generally confirm sex abuse are attractive to prosecutors, who can rely on them to provide the "definitive medical evidence," that is, the "proof" that sex abuse indeed took place.

Those who rarely find sex abuse are likely to be engaged by defense attorneys who invite them to testify that the child is "normal" and that there was "no evidence for sex abuse." Although there are people who claim that they are completely neutral, my experience has been that most people who are doing this kind of work have a reputation (whether warranted or not) for being in either of the two camps.

There are doctors (even pediatricians) who claim that any inflammation of a little girl's vulva is a manifestation of sex abuse. Most, however, note that this is an extremely common finding and can result from sweat, tight pants, certain kinds of soap, and the occasional mild rubbing (sometimes masturbatory) activity of the normal girl.

There are some who maintain that the normal hymen is a perfect circle (or close to it) without any irregularities. It follows, then, that if any irregularities are found, these must have been artificially created by the insertion of something, possibly a finger, possibly a penis, or possibly something else (like a crayon or pencil). There are others who claim that the normal hymen is most often not a circle and there are irregularities, tags, and bumps. They believe that these irregularities (sometimes referred to as serrated hymenal orifices) are within the normal range of hymenal variation.

Some claim that a three-year-old girl's vagina can accommodate an adult's fingers and even penis without necessarily showing signs of physical trauma, other than the production of the aforementioned irregularities, tags, and bumps. Others maintain that the insertion of an adult male penis into a three-year-old girl's vagina will produce severe pain, significant bleeding, and deep lacerations, and that the insertion of crayons and pencils at that age is extremely rare because of the pain and trauma that such insertion will produce.

There are significant differences of opinion regarding what is the normal size of the hymenal opening, and this, of course, bears directly on the question of abuse. Most experts agree that there have not been large studies of many children at different ages with regard to what the normal hymen looks like, its size, and whether or not it is indeed circular.

Furthermore, all agree that the older the child, the greater the likelihood the vaginal opening will accommodate a penis without significant trauma. Thus, by the age of nine or ten, one does not get the same degree of trauma that is found at younger ages.

Most agree, as well, that children of nine and ten, whose vaginal orifices are still small, could still be brought to the point of intercourse with an adult by gradual stretching of the vagina in the course of repeated experiences in which progressively larger objects (fingers, and ultimately a penis) are inserted.

Some physicians believe that a certain type of dilatation ("winking") of the anal mucosa is pathognomonic of penile penetration into the anus. There are others who claim that such dilatation is normal. (Here I am with the group that holds that such puckering is most often normal and is not a manifestation of sex abuse.)

The net result of this situation is that there may be sharply divided opinions among physicians regarding whether a particular child has been sexually abused. However, this does not stop each side from bringing in a parade of adversary physicians who predictably provide the "proof" that the child was sexually abused or that there is "no evidence" of sexual abuse. Another result of this situation is that many doctors are making a lot of money, because providing court testimony can be quite remunerative.

Definition of Terms
I will focus here on several terms that are often seen in reports of physical examinations of children being evaluated for sex abuse. Because girls are much more frequently subjected to sex abuse than boys, and because controversies regarding the signs of sex abuse are much greater in girls than boys, most of these comments relate to the physical examination of girls. It is assumed that the reader has a basic familiarity with the female genitalia and is familiar with such terms as labia majora, labia minora, clitoris, urethral meatus (orifice), hymenal orifice, and vaginal walls.

Examination Positions
Most often there are two positions described for a girl's examination, the supine frog-leg position and the prone knee-chest position. When examined in the supine frog-leg position, the child is on her back with her legs spread apart in "frog-leg" fashion. In the prone knee-chest position, the child's abdomen is close to the table and she is supported by her knees and chest.

McCann (1988) emphasizes the importance of the child's chest touching the table and the child's back being in a relaxed position. Examination of the vagina and cervix (without the use of a speculum) is more easily accomplished in young children in the prone knee-chest position. Sometimes a third position is utilized, the supine knee-chest position. Here the child lies on her back, puts her legs together, flexes her thighs at her hips, and is asked to hug her knees to her chest.

Hymenal Configurations
There are a wide variety of hymenal orifices and configurations. So great is their variation that some orifices do not easily lend themselves into being categorized. Furthermore, there is no strict standardization with regard to the names of the various kinds of openings. Accordingly, different examiners may use different names for the same hymenal configuration. The way in which the child is positioned may affect the hymenal configuration and thereby affect the name used by the examiner. I describe here the most common types of vaginal orifices. Next to each name I have placed in parentheses other terms that are often used for the same configuration.

Annular (Circumferential, Cuff-like, Central) This is the simplest configuration. The hymenal orifice is represented by a relatively even circle. Basically, it is a circular hole that can vary in diameter from almost a pinpoint to an enlarged orifice that leaves practically no hymen at all, only a rim. The cuff-like configuration is also annular, yet there is a thickening at the circumference of the orifice. Most competent examiners agree that the perfect circle type of hymen is not common.

Crescentic (Horseshoe, U-Shaped, Posterior Rim, Semilunar) The hymenal orifice is represented by a half-moon or crescent. The bottom of the U-shape, however, is at the posterior position (closest to the anus). The hymenal tissue, then, can appear as if it were hanging down from above (the anterior position).

Redundant (Denticular, Folded, Fimbriated, Serrated) Here the configuration is one in which tooth-like (denticular) tags of varying size project into the hymenal orifice. When these are relatively small, they give a saw-tooth (serrated) appearance. Because they are directed inward from the hymenal rim, they are called fimbriated (fringed).

The redundant configuration is quite common. The hymenal tissue projections are commonly referred to as tags and bumps. Estrogen has the effect of thickening the hymen and increasing the formation of these redundant projections into the hymenal orifice. Accordingly, the prepubertal girl is likely to have more such redundancies than younger girls. The spaces between these projections are often referred to as notches and clefts. These are to be differentiated from tears and lacerations, which suggest the insertion (partial or complete) of some object (animate or inanimate) beyond the hymen into the vaginal cavity. Whereas notches and clefts do not extend outward to the base (or periphery) of the hymen, tears and lacerations frequently do. And this is one of the important differentiating criteria between them. Furthermore, notches and clefts have rounded edges, whereas tears and lacerations have sharp edges.

Vascularity of the tissue around notches and clefts is even, smooth, and continuous with the vascularity and color of the rest of the hymen. Just as estrogen increases redundancy, it also has the effect of thickening the hymen and obscuring thereby the fine lacy vascular pattern typical of younger girls. This thickening also results in a loss of the translucency of the hymenal tissues, and the thickening gives the appearance of rounding of the edge of the hymenal membrane. Tears and lacerations are surrounded by tissue of different color, depending upon the period between the trauma and the time of the examination. The terms healed tears and scars are used to refer to stages of healing. I will comment further on these terms in the sections below.

Septate A septum is a partition or a dividing wall between two spaces or cavities. A septate hymen with one or more partitions (usually vertical) will result in two or more parallel (but also vertical) orifices.

Slit-like The hymenal orifice is represented by a thin slit, almost completely occluding communication between the vagina and the exterior.

Punctate (Cribriform) In this configuration there are multiple extremely small (pinpoint) orifices.

Imperforate Here there is no hymenal orifice at all. This may not cause difficulties prior to puberty. After the child stats menstruating, however, incision of the hymen is necessary if there is to be proper release of the menstrual flow.

The term anterior is used to refer to that past of the hymen that is closest to the front of the body, and the term posterior to that part of the hymen that is closest to the back of the child's body. Commonly, the site of a particular observation is described by visualizing the hymenal ring to be like the face of a clock. Accordingly, 12:00 o'clock would be the most anterior position; 3:00 o'clock the position closest to the child's left side (the examiner's right); 6:00 o'clock, the position closest to the child's anus; and 9:00 o'clock the position closest to the child's right side (the examiner's left). There is a widespread belief that attempts to insert an object (animate or inanimate) into the child's vagina is more likely to produce trauma to the posterior rim of the hymen, namely, in the range from the 3:00 to 9:00 o'clock position.

Sometimes examination of the hymen may be compromised by the presence of labial adhesions. These cause a sticking together of adjacent parts of the labia minora. Sometimes the attachment is by fibrous bands, and sometimes merely by a sticking together of labial tissue. These are so common that they are generally considered to be in the normal range. Most competent examiners would not consider them, per se, to be a sign of sex abuse.

Additional Terms
Here I define further terms frequently seen in reports by examiners assessing for sex abuse.

Labial adhesions This term refers to the "sticking together" of the labia minora and/or labia majora. Other names for the same phenomenon include labial agglutination, vulvar fusion, vulvar synechiae, gynatresia, coalescence of the labia minora, and occlusion of the vaginal vestibule. Labial adhesions are usually seen between the ages of two months and seven years. They are generally considered to be the result of poor hygiene, a mild vulvitis, or mechanical irritation along with hypoestrogenism (McCann, Voris, & Simon, 1988).

Synechiae This refers to a pathological union of parts. It is synonymous with the word adhesion. It is best viewed as a sticking together of parts that should be separate from one another. Infection and irritation can cause synechiae.

Posterior fourchette A fold of mucous membrane just inside the point of posterior conversion of the vulva (labia majora).

Examining Instruments
Because the hymenal structures are so small (the average normal hymenal orifice of a three-year-old is 4-5 mm) and because measurements may be difficult, variable, and somewhat subjective, visualization aids are often used. One such aid is the traditional otoscope. Although designed for examination of the ears, it has proven useful in the genital examination as well. It is basically a flashlight with a cone-shaped attachment and magnifying glass that, at the same time, focuses a beam of light on the area to be examined and allows the examiner to have a magnified view of what is being seen.

A superior instrument is the colposcope. The colposcope is a pair of mounted binoculars which can be mounted on a tripod or suspended from a movable mechanical arm. It generally magnifies from 10 to 20 times. The colposcope allows for visualization of structures that may not be visible to the naked eye. The colposcope is also equipped with an internal light for better visualization. It includes a green filter that assists in the examination of the vascular bed. Finkel (1989) states: "The green light improves visualization of scar tissue and alterations in the vascular pattern of the hymenal membrane and perihymenal tissues." Special cameras can be used to take photographs through the colposcope. The terms colposcopy and colposcopic examination refer to the procedure in which the colposcope is used. McCann (1990) has written an excellent description of the colposcope and its use. Muram and Elias (1989) have reservations about the colposcopy and do not consider it significantly superior to the unaided eye.

The vaginal speculum is an instrument that allows for visualization of the vaginal wall and the cervix. It is best visualized as a split tube with a special handle. The tube is inserted into the vagina and by squeezing the handle the tube expands, thereby widening the vagina and allowing for visualization of the cervix and vaginal wall, especially while the speculum is being removed. Although it comes in various sizes, it is rarely used in the examination of children. The insertion of a vaginal speculum into the vagina of a child would be very painful, and even traumatic, especially to the hymenal ring.

The Tanner Stages
The Tanner stages are used to describe objectively the developmental level of the secondary sexual characteristics in children and adults. The stage levels are divided into three categories: breast, genitals, and pubic hair. For each of these there are five or six stages, ranging from the most immature to the most mature. For example, Stage I of pubic hair development is no pubic hair at all. Stage II of breast development is the presence of a breast bud, with elevation of the breast and nipple on a small mound. Stage V of genital development in the male is a penis of adult size and shape. Although the Tanner stage has little if anything to do with sex abuse, the term is frequently seen in the medical reports of children being evaluated for sex abuse.

The [Examination?]
Although the physical examination in cases of suspected sex abuse will not be discussed in great detail here, there are some important areas to consider in evaluating the significance of such an examination.

According to Muram (1989a), it is important for the examiner to examine the child within one week of the alleged assault. It is in that period that residual bruises and inflammation are more likely to be present. Beyond that time these associated findings are likely to disappear. The time between the alleged assault and the examination should be noted in the report.

A common practice is for the physician who conducts the examination to form conclusions about sex abuse purely on the basis of the physical examination. The justification is that others should be responsible for delving into the background information, which can shed light on whether the sex abuse did indeed take place.

The doctor may claim, "I'm a doctor, not a detective. My job is to describe medical findings; others concern themselves with the investigation." I do not agree with this position. When examining for the presence of other diseases, that same doctor would certainly ask questions of one or both parents in order to obtain a "history" and thereby get more information about the disease under consideration.

Like most things in life, there is a continuum from the zero-to-hundred level of involvement. A physician who only is concerned with the physical examination is at the zero level in terms of getting historical background information. Most physicians who examine for sex abuse will go a little beyond that and get some information from the party who brings the child, most often the mother. Usually, such data collection does not occupy more than a minute or two. Accordingly, there is little meaningful inquiry into the details of the allegation and little opportunity to assess its credibility and likelihood. I have never (I repeat never) seen a medical report in which the examiner has seen fit to invite the alleged perpetrator (even when the person accused is the father — the most common case) to provide input.

Most often the examiner will state that the findings are "consistent with sex abuse." However, I have seen reports in which the alleged perpetrator is named, even though that party was not only not seen but there wasn't even an invitation extended to provide information. Such a practice is unconscionable and is worthy, in my opinion, of a malpractice suit. Such a physician is basically making a diagnosis on a person whom he or she has never seen. I am certain that the same doctor would be very reluctant to write any other diagnosis in a chart regarding a person who was not directly examined.

The failure to get information from available alleged perpetrators has caused much unnecessary grief. I cannot criticize such physicians strongly enough. Although state laws generally require the physician to report suspected abuse, they do not prevent the physician from speaking with the alleged perpetrator before making a final decision regarding whether a referral and investigation are warranted. Furthermore, many of these physicians do not appreciate the degree of ineptitude of the "validators" to whom they are referring their patients. They seem to be operating under the delusion that these people are competent in the area of differentiating between true and false sex abuse accusations.

As physicians they are sworn to subscribe to the Hippocratic oath in which they vow that they will "above all do no harm" to their patients. There is no question that many of the children who are referred to child protection services, evaluated by "validators," and others of that ilk are being seriously traumatized and that the physician has played a role in contributing to such trauma. I am not suggesting that physicians break the law. I am only suggesting that they take the time to get more information before making such referrals. I am also pointing out the common ineptitude of those people to whom they are referring their patient for the "final decision."

Physicians must also appreciate how their "impressions" and statements (for example, "consistent with sex abuse"), although not conclusive in their minds, are interpreted by many lay people as the final "proof." In many cases "consistent with sexual abuse" becomes transformed into "physical evidence of sexual abuse." Perhaps if physicians appreciated this more, they would be less quick to come to conclusions.

The measurement of the hymenal orifice is considered an important part of the physical examination of girls suspected of being sexually abused. It is important to appreciate how variable this finding can be. It differs according to the examination technique used (McCann, Voris, Simon, & Wells, 1990). Yet, there are people who are in jail because of this one measurement. The horizontal (transverse) diameter of the hymenal orifice is usually measured in the supine frog-leg position. Many factors are operative in determining what this diameter is. If the child is correctly positioned, the heels will be placed just below the buttocks. Clearly, if they are in another position, such as 12 inches below the buttocks, a different measurement will be obtained.

The examiner must be sure that the child's heels are at the same position assumed by those children on whom the normative data were obtained. Then there is the variable of the degree to which the child's legs are spread. Usually, an assistant stands next to the child and slowly spreads the child's legs while distracting and reassuring the child. Obviously, the greater the degree of spread, the wider will be the hymenal orifice. However, even when the legs are extended to the most extreme position that is comfortable, the labia majora are usually still so close to one another that the hymen will not be observable. Accordingly, the assistant generally pulls the labia majora apart laterally and posteriorly in order to allow hymenal visualization.

Obviously, there are varying degrees of such posterolateral traction, and the greater the traction, the greater the expansion of the hymenal orifice. Therefore, the assistant must attempt to apply such traction to the same degree applied by those collecting the normative data. A common standard is for the assistant to apply traction at the mid-point of the labia majora to a point 1-1.5 cm on either side of the midline.

Furthermore, a lag must be allowed between the time of retraction and the time of taking the measurement. There is usually a 1-2-second period during which the hymenal ring must be allowed to dilate. Competent examiners usually allow at least a 3-4-second time lag in order to ensure that the hymenal ring is going to relax into its resting position. McCann (1988) and McCann, Voris, Simon, & Wells (1990) emphasize that the greater the traction on the labia majora, the greater the width the hymenal diameter will be, and this is one of the explanations for why different examiners get different results when measuring hymenal openings. They also point out that the vertical diameter is smaller in the supine frog-leg position than it is in the prone knee-chest position.

A small millimeter ruler is then placed very close to the vaginal opening. Obviously, any squirming by the child is going to compromise the accuracy of this measurement. However, even under optimum conditions, and even with strict reproduction of the positioning used by those collecting the normative data, there is bound to be some variability of measurement because of the minuteness of the measurement being considered here. A millimeter is approximately 1/25 of an inch. Although the human eye is capable of discriminating between, let us say, 4 mm and 5 mm, it is obvious we are dealing here with a discrimination that is close to the edge of the capability of the human eye (and brain). One has to consider also that the distance of the examiner's eye from the hymenal orifice and the distance of the ruler from the hymenal orifice can very well affect the measurement perceived by the examiner.

I am convinced that if the same examiner were to examine the same child on the following day, even when attempting to reproduce exactly the conditions of the examination, there would be variability. Furthermore, another examiner, again under the same circumstances, is also likely to come up with a different measurement. The American Academy of Pediatrics (1991) in its statement, "Guidelines for the Evaluation of Sexual Abuse of Children," emphasizes the aforementioned variability and impresses upon pediatricians the importance of taking these variations into consideration when making decisions regarding the normality or abnormality of the size of the hymenal orifice.

The prone knee-chest position is generally used to measure the vertical diameter of the hymen. Here, too, lateral traction is required if one is to properly visualize the hymen and there is great variability regarding the child's positioning and the degree of lateral traction. Again, standardization is necessary. McCann (1990) states: "The head is turned to one side with the forearms resting on either side of the head. The knees are separated 6-8 inches and maintained in 90 degrees of flexion. The examiner's thumbs are then placed beneath the leading edge of the gluteous maximus at the level of the vaginal introitus and the posterior portion of the perineum is lifted, revealing the hymenal orifice."

Obviously, the examiner who does not follow this procedure exactly will obtain different measurements of the hymenal orifice. Examination in the prone knee-chest position allows the hymenal tissues to fall forward and thereby provides better visualization of the full circumference of the hymenal orifice than is generally possible in the frog-leg position. Horowitz (1987) provides a good general statement of procedures for conducting a pediatric examination for sex abuse, as does the American Academy of Pediatrics, Committee on Child Abuse and Neglect (1991).

What are Normal Genital Medical Findings?
Female Genital Findings
As mentioned above, it has only been in recent years that extensive studies have been done to determine normal genital findings in children. This belated interest relates to the rapid increase in reports of sex abuse and the need for accurate data in order to differentiate the normal from the sexually abused child. It is my hope that the reader will now be impressed with the complexity of the problem of obtaining normative data with regard to the hymenal orifice, and will be even more overwhelmed by the complexity of the problem after a discussion of the wide variety of seemingly pathological configurations that are found in normal children,

First, with regard to data collection on the size of the normal hymenal orifice, one of the problems attendant to conducting such studies is that of knowing with certainty that the children studied were not abused. It is impossible to "prove" that "something didn't happen." The greater the number of children included in a study, the greater the likelihood the findings will be credible. However, the greater the number of such child subjects, the less the likelihood that each of them was studied in depth with regard to whether or not they were sexually abused.

The fact that children were taken from a "normal population" of youngsters who were not referred for abuse is no guarantee that some of the subjects being studied were not abused. This is one of the criticisms directed at such studies, especially by those who tend to diagnose sex abuse in the vast majority of patients refereed to them. These individuals are likely to use as criteria findings that other observers would consider to be in the normal range. This is one of the major problems in this field, and it is a significant source of controversy.

Goff, Burke, Rickenback, and Buebendorf (1989) studied 273 prepubertal girls as part of their routine health assessment. They measured horizontal diameters only in the supine knee-chest position and the supine frog-leg position. No measurements were made in the prone knee-chest position. The girls ranged in age from under age 1 to age 7. This study, as is true of most studies, confirmed that the vaginal orifice increases in size with age. The authors found that the horizontal hymenal diameter was generally larger when measured in the supine knee-chest position than in the supine frog-leg position. Interestingly, an orifice greater than 4 mm in horizontal diameter was rare. The study is a very good one, especially because the authors describe in great detail the exact positioning of the children prior to measurement.

McCann, Wells, Simon, and Voris (1990) studied 93 girls between the ages of 10 months and 10 years. Whereas Goff et al. (1989) used direct visual measurements, McCann et al. (1990) used a colposcope. McCann et al. took both vertical and horizontal measurements in the supine position with labial separation, the supine position with labial traction, and the prone knee-chest position. McCann et al's findings are different from those of Goff et al., in that the hymenal orifices were typically larger. There was only one mean measurement below 4.0 mm, and that was the horizontal measurement in the supine labial separation position, namely, 3.9 + 1.4 mm. The largest finding was for the 8-year-old girls in the 8-l0-year group in the prone knee-chest position, namely, the vertical diameter of 8.7+2.6 mm. Considering these extremes, one can see that the range of the means goes from 3.9 to 8.7 mm.

Accordingly, physicians who believe that any measurement over 4 mm is indicative of sex abuse (which would be suggested by Goff et al.'s studies) would not find support in McCann et al.'s studies. Both are competent examining teams and both have written articles that are very impressive. Yet, they would be quoted by adversaries in a courtroom dispute regarding whether or not sex abuse took place.

Finkel (1989) holds that a transverse hymenal diameter of greater than 5 mm is suggestive of sexual abuse. However, because of the unreliability of such measurements, repeated measurements must be taken before coming to a conclusion. He also emphasizes that the position of the child and the degree of relaxation are important factors in determining the measurement.

Another reliable study was conducted by White, Ingram, and Lyna (1989). Their subjects were 242 females, ages 1-12. Three groups were studied: (1) sexually abused, (2) no history of sexual contact, but at risk, (3) non-abused. Transverse diameters only were obtained with patients in the supine frog-leg position. Lateral tension was applied to the hymenal opening. Measurements were made by visualization of a measuring tape held over the hymenal orifice or by a cotton-tipped applicator. They found that 88% of children who complained of penile/vaginal penetration had a vaginal introital diameter of greater than 4 mm, as compared to 18% of children who described no such penetration. They concluded that a vaginal introital diameter of greater than 4 mm is highly associated with sexual contact in children less than 13 years of age.

It is important to appreciate that the transverse diameter of the average adult erect penis is approximately 3.5 cm (35 mm) and an index finger is approximately 1.5 cm (15 mm) wide. Accordingly, the insertion of either of these into a hymenal orifice of 5 mm will invariably cause significant widening and, certainly in the younger girl, pain and trauma. Accordingly, when a three-year-od girl claims that an alleged perpetrator inserted his penis into her vagina and the vaginal examination reveals a diameter of, for example, 7-8 mm, it is extremely unlikely that the penetration being described actually took place. The more likely explanation is either examiner error or the hymenal orifice is at the upper end of the normal bell-shaped curve of hymenal diameters.

McCann, Wells, Simon, and Voris (1990) describe other observations relevant to the problem of differentiating the non-abused from the sexually abused children. For example, some claim that rolled hymenal edges are a manifestation of sex abuse. However, McCann et al. found that the rolled edge is much more commonly seen in the supine positions, but tends to disappear in the knee-chest position. Finkel (1989), in contrast, states that rounded hymenal edges are one of the results of the effects of estrogen in the prepubertal girl and are more likely to be visualized in the knee-chest position. This not only says something about the importance of positioning, but also says something about rolled edges as a sign of sex abuse.

With regard to hymenal configuration, McCann et al. (1990) found

• crescent (36%),
• concentric [annular] (32%),
• septate (1%),
• cribriform (0%),
• imperforate (2%).

In 17% of the subjects he was unable to determine the exact configuration because of redundancy of hymenal tissues and the failure of the hymenal orifice to open. These findings lend confirmation to those who claim that a perfectly circular hymen is not the only configuration. With regard to the hymenal edge, he found the following:

• smooth (26%),
• irregular (25%),
• redundant (25%), and
• angular (8%).

Again, these findings lend support to those who hold that there is great variation in the configuration of the hymenal orifice. In the traction frog-leg position, with regard to some of the "abnormalities" sometimes considered manifestations of sex abuse, he found the following:

• thickened hymenal edge (53.8%),
• localized roll of the hymenal edge (23.8%),
• hymenal mounds (33.8%),
• hymenal projections (33.3%),
• hymenal tags (24.4%),
• peri-hymenal bands (16%),
• septal remnants (8.6%),
• hymenal septa (2.5%),
• hymenal notches (6.6%),
• hymenal synechiae [adhesion of the hymen to adjacent tissues] (2.4%).

Some claim that the normal hymen is regular in its vascularity and any areas of vascular irregularity, areas in which the vascularization is different from surrounding tissues, is strongly suggestive of healed tears and other signs of sex abuse.

McCann et al. (1990) found irregular vascularity in

• 31.3% of those children examined in the separation frog-leg position,
• 30.9% in those children examined in the traction frog-leg position, and
• 28.9% of those when examined in the knee-chest position.

Aside from areas of irregular vascularity, they found areas of isolated increase in vascularity in

• 13.9% of those examined in the separation frog-leg position,
• 16.0% of those examined in the traction frog-leg position, and
• 22.8% of those examined in the prone knee-chest position.

These findings strongly suggest that the vascular irregularity criterion for sex abuse is improper and risky (especially for those being falsely accused).

The McCann et al. study directs itself, as well, to the frequency of other "abnormalities" sometimes considered manifestations of sex abuse.

For example,

he found labial adhesions to be present in 38.9% and periurethral bands in 50.6% of the children studied.
He found erythema of the vestibule to be present in 56% of the children examined.

(The vestibule is the portion of the vulva bounded by the labia minora. At the floor of the vestibule are [from anterior to posterior] the clitoris, urethral orifice, and the hymen.)

As mentioned previously, vulval rashes are quite common in children. These relate to poor hygiene, a wide variety of infections (not necessarily related to sexually transmitted diseases), tight panties, certain soaps, rubbing, scratching, and masturbation (to mention the most common). I have been involved in a number of cases in which these more common and likely causes of the erythema were ignored and the examiner concluded that the findings were "consistent" with sex abuse or even manifestations of sex abuse.

I have discussed in some detail the McCann, Wells, Simon, and Voris (1990) research because it provides compelling evidence that normal children exhibit a wide variety of variations, many of which have been considered signs of sex abuse. It is of interest that McCann et al.'s original group consisted of 114 girls, but 23 were excluded because of the early onset of puberty and the possibility of undetected sexual abuse.

The list of behavioral manifestations that warranted their exclusion from the study included nightmares, fears, moodiness, change in school performance, truancy, and acting out behaviors (among others). All of these could be seen in normal children (at least on occasion), and many of these behaviors are manifestations of a wide variety of childhood problems completely unrelated to sex abuse. There are sexually abused children, however, who may exhibit one or more of these behavioral manifestations.

To the best of my knowledge, McCann et al. did not conduct a detailed inquiry regarding whether these behavioral manifestations were signs and symptoms of sex abuse, were in the normal range, or related to other causes. On the one hand, the exclusion of all these children, simply on the basis of the presence of one or more of these symptoms, made his sample "purer" — thereby lessening the likelihood that sexually abused children were included. On the other hand, he may have unnecessarily shrunk his patient population, thereby lessening somewhat the credibility of his findings and depriving himself of many subjects who were not molested.

Anal Findings (Male and Female)
Anal and perianal findings are also a source of significant controversy. One of the most widely known such controversies relates to the anal examinations described by Hobbs and Wynne (1986, 1987). These examiners claim that a pathognomonic sign of child sex abuse is "reflex dilatation and alternate contraction and relaxation of the anal sphincter or 'twitchiness' without dilatation." One finding, also referred to as anal "winking," is considered a pathognomonic sign of anal intercourse. As a result of using this criterion, hundreds of children in England were diagnosed as having been sexually abused, with the result that 121 children were removed from 57 families. It took a government investigation to bring society to its senses and return these children to their families.

McCann, Voris, Simon, and Wells (1989) studied 267 children (161 girls and 106 boys), ages 2 months to 11 years. They found anal dilatation in 49% of the children, and the mean time of the initial dilatation was 65 seconds. The anus opened and closed intermittently in 62% of the subjects in which dilatation occurred. Accordingly, about 30% of all the children studied exhibited the intermittent dilatation and relaxation of the anal sphincter, which Hobbs and Wynne considered a sign of sex abuse.

McCann et al. (1989) describe other anal findings in normal children that are often considered signs of sex abuse. They found that 41% of their group exhibited erythema. There is no question that children who have been sexually molested per anus will exhibit erythema. But in this study, 41% of normal children exhibited erythema as well. McCann et al. found increased pigmentation in 30%, another finding that is often considered a sign of sex abuse. They found venous engorgement in 52% after two minutes in the knee-chest position. Again, venous engorgement has also been considered a sign of sex abuse. Anal tags and folds are also considered by some to be indicative of sex abuse. These were found anterior to the anus in 11% of the children studied. No abrasions, hematomas, or fissures (common findings in sex abuse) were found.

What are the Genital Findings in Sexually Abused Children?
Studies of the anogenital findings in sex abuse are beset by a number of problems. First, all knowledgeable investigators agree that some children who have been genuinely abused sexually will exhibit no medical findings. This relates to the fact that they were caressed and touched in a way that would not be expected to cause physical trauma. Another problem relates to the fact that the investigators can never be sure that all the children in the non-abused group studied were indeed never abused. There is also the risk that some of the children in the abused group were indeed not abused, but this is less likely. A third problem relates to the fact that a wide variety of abnormalities are seen in normal children, and the aforementioned studies of McCann and his colleagues provide good verification of this. What we are trying to find, then, are specific medical findings that are seen only in abused children and not in those who have not been abused.

Female Genital Findings
Emans, Woods, and Flagg (1987) studied 305 girls. They were divided into three groups:

(1) sexually abused (119 girls),
(2) normal girls with no genital complaints (127 girls), and
(3) girls with other genital complaints (59 girls).

The abused group was more likely to have

scars on the hymen or the posterior fourchette (9% vs. 1%, p < 0.002),
increased friability (ease of bleeding) of the posterior fourchette (10% vs. 1%, p < 0.001),
attenuation (stretching and thinning) of the hymen (18% vs. 4%, p < 0.0003), and
synechiae (adhesions) from the hymenal ring to the vagina (8% vs. 0%, p < 0.0009).

We see here that we are not dealing with a situation in which a finding is present in the abused group and not present in the non-abused. Rather, certain findings are more likely to be present in the abused group than in the non-abused group. The obvious problem with this kind of finding is that its presence then does not necessarily mean that the particular child being examined was abused.

Interestingly, Emans et al. (1987) found a wide variety of symptoms to be present with equal likelihood in the abused group and the non-abused group with other genital complaints. There was no statistical difference between groups 1 and 3 regarding the frequency of abrasions, hymenal tears, intravaginal synechiae, and condyloma acuminata (venereal warts). This study, then, suggests that these particular findings are not of diagnostic significance when attempting to differentiate abused from non-abused children.

Interestingly, erythema (reddening) was more common in the non-abused group than in the abused group (68% vs. 34%, p < 0.0001). There was no statistical difference between the dimensions of the hymenal opening of the abused and the non-abused group. One would certainly expect a larger average hymenal opening in the abused group, but this study did not confirm such a difference. Perhaps there were too few girls in the 119 abused who had the kind of sexual molestation that would produce an enlargement of the hymenal ring.

However, as Herman-Giddens and Frothingham (1987) point out,

"The hymen, contrary to common notion, is often a slack, thick, folded, stretchable tissue which may persist after digital or penile penetration."

The same authors hold that

"a vaginal opening of greater than 5 mm is not common and may indicate vaginal penetration with a finger, object, or penis."

McCann (1988) states that 85% of preadolescent children who are being molested are molested on a chronic, ongoing, and recurring basis. Such molestation should, then, produce changes indicative of chronic trauma. He emphasizes the importance of examination for bruises in other parts of the body, in the nongenital area. The mouth is a common site of lesions because the perpetrator may have placed his hand over the child's mouth in order to stop the child from screaming. Grab marks on the arms and inner thighs are also strongly suggestive of sex abuse, especially thumb marks on the inner aspect of the thigh, placed there when the child's legs were forced apart.

McCann (1988) also observes that labial injury is common at the time of rape because the labia majora are generally closed and the perpetrator pushes his penis repeatedly against closed labia. He believes that the most common area of hymenal injury is between the 4:00 and 7:00 o'clock positions because the penis is forced downward and backward. He emphasizes that children heal quickly and that examinations after the first few days may not confirm the abuse. Because the length of the vagina of four- and five-year-old girls is only 4 cm, trauma to the vagina, cervix, and lower part of the uterus is common.

McCann, Voris, and Simon (1988) studied six sisters, all of whom had been sexually molested by male family members. All of these girls had labial adhesions, and four of the six had changes in the area of the posterior fourchette (a fold of mucus membrane just inside the posterior commissure of the vulva). Furthermore, four of the girls' hymens revealed abnormalities of the hymenal edge (irregular, rolled, or septum) and three revealed irregularities of the hymenal membrane (redundant, thick, scarred). Four exhibited abnormal vascular patterns, and all six exhibited adhesions and/or scars of the posterior fourchette. The labial adhesions in these cases were associated with posterior fourchette changes and other findings consistent with sex abuse.

The authors' position is that labial adhesions per se are not indicative of sex abuse. However, if associated with other findings suggestive of sex abuse, such as posterior fourchette trauma, then it should be considered one such manifestation. We see here, then, a situation in which a normal finding is considered a sign of sex abuse under certain circumstances. In these six cases the labial adhesions were associated with other findings indicative of sex abuse. Furthermore, labial adhesions usually occur from ages two to seven. In this case two of the girls were ages eight and nine, beyond the age at which one usually sees labial adhesions.

Muram (1989a) divides the genital findings into four categories:

1. Normal-appearing genitalia.
2. Nonspecific findings.
   Abnormalities of the genitalia that could have been caused by sexual abuse, but also are often seen in girls who are not victims of sexual abuse (e.g., inflammation and scratching). These findings may be the sequelae of poor perineal hygiene or nonspecific infection. Included in this category are redness of the external genitalia, increased vascular pattern of the vestibular and labial mucosa, presence of purulent discharge from the vagina, small skin fissures or lacerations in the area of the posterior fourchette, and agglutination of the labia minora.
3. Specific findings.
   The presence of one or more abnormalities strongly suggesting sexual abuse. Such findings include recent or healed lacerations of the hymen and vaginal mucosa, enlarged hymenal opening of 1 cm, proctoepisiotomy (a laceration of the vaginal mucosa extending through the rectovaginal septum to involve the rectal mucosa), and indentations in the skin indicating teeth (bite) marks. This category also includes patients with laboratory confirmation of a venereal disease.
4. Definitive findings.
   Any presence of sperm.

It is of interest that Muram (1989a) considers labial agglutination to be a nonspecific finding, in that it does not necessarily indicate sex abuse. Of importance in the third category, specific findings, are hymenal tears that extend to the base of the hymenal ring as to be differentiated from hymenal clefts which do not extend that peripherally.

Muram believes that an astute examiner will do just as well with the unaided eye as with the colposcope. Muram (1989b) studied 31 girls who were assaulted by 30 individuals, all of whom confessed to having sexually molested them. Both the girls and the perpetrators were in agreement that the sex abuse took place. Obviously, this is a good study sample for ascertaining the physical effects of sex abuse. It circumvents one of the aforementioned problems regarding such studies, namely, the uncertainty regarding whether or not the girl being examined was genuinely abused or was genuinely in the non-abused category.

In 18 of the 31 cases the offender admitted to vaginal penetration. However, specific findings were only to be found in 11 of these 18 girls (61%). In those girls in which penetration was denied only 3 of 13 (23%) provided specific findings. However, the girls ranged in age from 2 to 15, so it is not surprising that some of the teenagers who experienced penile penetration did not have physical findings of abuse.

It is of Interest that of the 31 girls, inflammation, bruising, and irritation were seen in only 9, all of whom were evaluated within one week of the assault. None of the girls evaluated one week after the abuse had findings suggestive of inflammation. Muram states: "If no tear of the hymen occurred, the examination will fail to detect any abnormalities." This is an important point. According to Muram, the most important specific sign of sexual molestation Is hymenal tear, to the base, especially extending into the vaginal canal. Other abnormalities, such as inflammation and bruising, tend to heal within a week.

The most important observation Muram makes is that the most consistent finding in bona fide sex abuse is laceration or tear of the hymenal ring, down through the base, and extending often into the adjacent vaginal wall. This sign is one of the most important for differentiating genuine from fabricated abuse.

On occasion, a child may sustain significant genital injuries associated with trauma to the perineal area as a result of falls and fence or straddle injuries (Behrman & Vaughn, 1983; Paul, 1986). Here one may see the kinds of lacerations seen in sexual abuse. One may also see abrasions and other forms of injury to the perivaginal area. However, the time of the trauma is generally well known to the child (and usually an adult), and there is nothing else in the history to suggest sexual abuse.

Paul (1977, 1986) claims that penile penetration in younger children will cause widespread injuries, including lacerations of the hymen, vagina, and labia. There will be profuse bleeding and the child will experience excruciating pain. This is an important point because in many cases of fabricated sex abuse, the child will describe no pain or minimal pain.

Anal Findings
McCann (1988) observes that children who have been subjected to anal intercourse on repeated occasions suffer with a relaxation of the external anal sphincter, but not of the internal anal sphincter. Accordingly, there is a typical funnel-like appearance of the anus on physical examination.

Finkel (1989) reports on seven children who had experienced acute genital and anal trauma in association with sexual abuse. Some of the more superficial manifestations of the trauma (abrasions, superficial lacerations, contusions, and bleeding) were not apparent after four days. In two of Finkel's seven cases, penile-anal penetration was involved. In one case, Finkel described "superficial lacerations of the anal verge tissues in anterior and posterior midline positions each measuring 2 mm circumferentially and 3 mm in length." In the second case he described five mucocutaneous superficial lacerations, some of which extended from the external anal mucosa down into the anal canal.

Paul (1990) observes that, even with the use of a lubricant, penile penetration of the anus will almost invariably result in some injury to the anal verge. He stresses the importance of the history, from the child, of severe pain — not only during the abuse, but when the child next attempts to have a bowel movement. He states: "This exacerbation of pain on defecation is an almost invariable 'story' and is so impressed on the child's mind that it is rarely forgotten" (p. 6).

Sexually Transmitted Diseases
The presence of a sexually transmitted disease (previously referred to as venereal disease) is generally considered definitive evidence for sex abuse. Of the wide variety of such diseases, the most commonly found in sexually abused children are gonorrhea, syphilis, Chlamydia, condyloma acuminatum, Trichomonas vaginalis, and herpes 1 (genital). However, it is important to appreciate that gonorrhea, syphilis, and Chlamydia can be acquired perinatally from the mother, and this must be given consideration before deciding that the presence of such a disease automatically indicates sex abuse (American Academy of Pediatrics, Committee of Child Abuse and Neglect, 1991).

The material for gonorrhea culture is generally obtained from cotton swabs of the vagina, throat, and rectum. The organism may sometimes be grown from cultures of the urine of suspected boys. The urine can also be examined for Trichomonas infection. Tests for syphilis are usually obtained from a blood sample. Vaginal secretions can also be cultured for the presence of Chlamydia, herpes, and Trichomonas. Vaginal secretions can be examined directly (microscopically, with proper staining) for gonorrhea and Trichomonas.

Condyloma acuminatum is also referred to as genital warts and venereal warts. It is caused by a virus called the human papilloma virus (HPV). It is the most common viral sexually transmitted disease in the United States and is now more common than herpes (due to the recent rapid increase in its incidence). Because the incubation period is approximately one month (Stewart, Stewart, Guest, & Hatcher, 1987), the genital warts will not be observable immediately after a child has been abused. The diagnosis is made generally by direct observation, the warts usually appearing like warts on other parts of the body, but they do extend into the vaginal canal, cervix, and rectum. Sometimes the warts are inconspicuous or completely invisible to the naked eye. Horowitz (1987) provides an excellent protocol for the examination for sexually transmitted diseases.

Although the presence of a sexually transmitted disease is strongly suggestive of sex abuse, the disease may have been acquired by the child in a nonsexual way. The problem in such situations is that the suspect may also have the sexually transmitted disease but did not have a sexual encounter with the child. Rather, the disease was transmitted nonsexually. Clearly, an accused who is trying to deny a sexual encounter will give strong support to this theory.

Support for this can be found in the medical literature, where there are many articles providing instances of just such a method of transmission. For example, Shore and Winklestein (1971) claim that 50% of their sample of children contracted their gonococcal infection in the absence of sex abuse and that only one-fifth acquired the gonorrhea through a sexual experience. Kaplan (1986) claims that the gonococcus can survive outside the human body for up to 24 hours and cites a 1929 study in which several newborns in the same hospital nursery were found to have gonococcal infections. It was believed that the organism was transferred with thermometers. Wakefield and Underwager (1988) refer to studies in which gonorrhea was found to have been transmitted nonsexually among peers, via close physical contact with infected adults or indirect contact through bedclothes or hands. They also refer to the work of DeJong et al. (1982), who report that venereal warts can be transmitted through close nonsexual contact, during delivery, and by sexual encounters.

Sperm in the Vagina and the Pregnancy Test
The presence of sperm in the vagina of a prepubertal child is obvious evidence for sex abuse. It is proof that a postpubertal male has sexually penetrated the prepubertal girl. The presence of sperm in the vagina of a postpubertal girl is not necessarily evidence of sex abuse, in that she may have voluntarily had sexual relations without in any way being abused.

Fresh sperm can be examined directly under the microscope. After 24 hours sperm may not be viable enough for such direct examination. Sperm may be visualized with Wood's light, under which it becomes fluorescent. These fluorescent "tear drops" shine dramatically in contrast to other vaginal secretions that are examined under Wood's light (McCann, 1988). The examiner must take care to question the parents regarding whether the child has taken a bath between the time of the alleged abuse and the time of the examination. Obviously, if the sperm has been washed out, the Wood's light test will not be positive. The sperm sample can also be tested for the presence of acid phosphatase, an enzyme that is secreted by the prostate gland and is to be found in the ejaculate. Acid phosphatase is not normally found in the vagina.

In association with the examination for sperm, one must consider the pregnancy test. Obviously, the pregnancy test is not viable for prepubertal children, although there are reports of pregnancy in girls as young as eight and many examiners will routinely do them for children of that age and above. Although conducting a pregnancy test on a prepubertal child may seem unnecessary and even absurd, it is not completely so. There are children who are capable of becoming pregnant who have exhibited few, if any, signs of sexual maturity. And this is where the Tanner level of sexual development may provide information regarding whether or not the child could indeed be pregnant. An eight- or nine-year-old, exhibiting Tanner II and III levels, may very well be capable of pregnancy.

Conclusions
Although physicians have been performing medical examinations and drawing conclusions about sex abuse, their conclusions have often been ill-considered and unsupported by empirical data. The recent research on the characteristics of the genitals of normal, non-abused children provides the baseline information needed to evaluate physical findings. This research indicates that many of the physical findings often claimed to indicate probable sexual abuse are found frequently in non-abused children. This research must be taken into account when evaluating reports of medical examinations of children in cases of suspected sex abuse.

References
American Academy of Pediatrics, Committee on Child Abuse and Neglect (1991). Guidelines for the evaluation of sexual abuse of children. Pediatrics, 87, 254-260.

Behrman, R. E., & Vaughan, V. C. (1983). Textbook of Pediatrics Philadelphia: W.B. Saunders Co.

DeJong, A. R., Weiss, J. C., & Brent, R. L. (1982). Condyloma acuminata in children. American Journal of Diseases of Children, 136, 704-706.

Emans, S. J., Wood:. B. R., Flagg, N. T., & Freeman, A. (1987). Genital findings in sexually abused, symptomatic and asymptomatic girls. Pediatrics, 79, 778-785.

Finkel, M. A. (1989). Anogenital trauma in sexually abed children. Pediatrics, 84, 317-322.

Goff, C. W., Burke, K. R., Rickenback, C., & Buebendorf, D. P. (1989). Vaginal opening measurement. American Journal of Diseases of Children, 143, 166-168.

Herman-Giddens, M. B., & Frothingham, T. B. (1987). Prepubertal female genitals: Examination for evidence of abuse. Pediatrics, 80, 203-208.

Hobbs, C. J. & Wynne, J. M. (1986). Buggery in childhood: A common syndrome of child abuse. Lancet, 2(8510), 792-796.

Hobbs, C. J., & Wynne, J. M. (1987). Child sexual abuse: An increasing rate of diagnosis. Lancet, 2(8563), 837-841.

Horowitz, D. A. (1987). Physical examination of sexually abused children and adolescents. Pediatrics in Review, 9(1), 25-29.

Kaplan, J. M. (1986). Pseudoabuse — the misdiagnosis of child abuse. Journal of Forensic Science, 31, 1420-1428.

McCann, J. (1988, January 21). Health science response to child maltreatment conference. San Diego, CA.

McCann, J. (1990). Use of the colposcope in childhood sexual abuse examinations. Medical Clinics of North America, 37(4), 863-880.

McCann, J., Voris, J., & Simon, M. (1988). Labial adhesions and posterior fourchette injuries in childhood sexual abuse. American Journal of Diseases of Children, 142, 659-662.

McCann, J., Voris, J., Simon, M., & Wells, R. (1989). Perianal findings in prepubertal children selected for nonabuse: A descriptive study. Child Abuse & Neglect, 13, 179-193.

McCann, J., Voris, J., Simon, M., & Wells, R. (1990). Comparison of genital examination techniques in prepubertal girls. Pediatrics, 85, 182-187.

McCann, J., Wells, R., Simon, M., & Voris, J. (1990). Genital findings in prepubertal girls selected for nonabuse: A descriptive study. Pediatrics, 86, 428-438.

Muram, D., & Elias, S. (1989). Child sexual abuse — genital tract findings in prepubertal girls, II. Comparison of colposcopic and unaided examinations. American Journal of Obstetrics and Gynecology, 160(2), 333-335.

Muram, D. (1989a). Child sexual abuse — genital tract findings in prepubertal girls, I. The unaided medical examination. American Journal of Obstetrics and Gynecology, 160(2), 328-333.

Muram, D. (1989b). Child sexual abuse: Relationship between sexual acts and genital findings. Child Abuse & Neglect, 13, 211-216.

Paul, D. M. (1977). The medical examination in sexual offences against children. Medical Science and the Law, 17, 81-88.

Paul. D. M. (1986). What really did happen to Baby Jane? The medical aspects of the investigation of alleged sexual abuse of children. Medical Science and the Law, 26, 85-102.

Paul, D. M. (1990). The pitfalls which may be encountered during an examination for signs of sexual abuse. Medical Science and the Law, 30, 3-11.

Shore, W. B.. & Winklestein, J. A. (1971). Nonvenereal transmission of gonococcal infections to children. The Journal of Pediatrics, 79, 661-663.

Stewart, F. H., Stewart, G. K., Guest, F., & Hatcher, R. A. (1987). Understanding Your Body: Every Woman's Guide to a Lifetime of Health (). Bantam Books, New York.

Wakefield, H., & Underwager, R. (1988). Accusations of Child Sexual Abuse C. C. Thomas Publisher: Springfield, IL.

White, S. T., Ingram, D. L., & Lyna, P. R. (1989). Vaginal diameter in the evaluation of sexual abuse. Child Abuse & Neglect, 13, 217-224.