RADIOIMMUNOASSAY OF TESTOSTERONE AND ESTROGENS IN BLOODSTAINS FOR THE PURPOSE OF SEX IDENTIFICATION

The capabilities and the limitations involved in applying commercially prepared Radioimmunoassay (RIA) Kits, for testosterone (T) and estrogen (E) , for the purpose of identifying the sexual origin of a bloodstain were examined. A total of forty whole blood samples were obtained from equal numbers of young healthy male and female volunteers. Bloodstains were prepared by absorbing these blood samples onto 2 X 2 cm squares of white cotton linen. Two separate studies were conducted using these stains. Both studies examined the fractional recovery of T and E when extracted with diethyl ether, the ability of the RIA method to detect T and E present in the extract at three bloodstain ages (2, 30, and 60 days after preparation) and whether a T:E ratio could be utilized as a basis for discriminating between male and female bloodstain samples. The difference between the studies lies in the prior knowledge of the sex of each individual blood sample. In the first study the sex of each blood sample was known, (open study), whereas in the second study the individual sex was not known but the number of each sex was known (single-blind study). A cross-reactivity study was performed to determine the direct interference of several commonly used synthetic steroids with the RIAs.

Generally the information obtained from these genetic markers which is useful for forensic purposes deals with population frequency data. Each antigen or isoenzyme detected in a single bloodstain will decrease the possible number of people that might have left a particular stain.
By using additional marker systems, the forensic scientist can reduce that probable percentage dramatically.
In some instances, such as those crimes in which there are no control bloodstains available for comparison, the blood grouping data alone are of little value to the investigators. It is occasionally possible, however, to obtain additional information from the stain concerning some physical characteristics of the individual who left the stain.
The possibilities of discerning the race, age and sex of the individual who left the stain are currently being investigated (Brown, 1977;Garner, 1979 andWigmore, 1979).
It is the determination of the sexual origin of a bloodstain through the quantification of t~e sex hormones: Testosterone (T) and Estrogens (E) by radioimmunoassay (RIA) techniques, that will be the subject of this study.
This study will attempt to investigate three major questions related to the RIA technique of determining the sexual origin of a bloodstain. The first question deals with the feasibility of whether commercially available kits are sensitive and selective enough to quantitate the amounts of each steroid present in a specific area of bloodstained material. The quantities that are obtained will then be calculated as a ratio of T:E and this final result will be used as a basis for discriminating between male and female samples in both open and blind studies. Secondly, this study will investigate the effect that the age of a bloodstain has on quantifying the steroids in stains up to 60 days old. Finally, the question of whether synthetic steroids, such as those used in oral contraceptive preparations, significantly cross-react in either RIA method will be examined.
Testosterone and estrogen are well suited to indicate sex based on the well established normal average values for each in males and females between 15 and 45 years old.
In a preliminary study only testosterone was assayed in stains prepared from whole blood obtained from ten male and six female volunteers. Although the results from this study demonstrated that sex determination is possible using only testosterone values, it also showed fluctuations in values between individuals of the same sex as well as some overlapping of values between sexes. The concurrent determination of total estrogens in a stain will hopefully stabilize variation between individuals. Also the use of the two markers in a ratio form to indicate sex will eliminate the need to estimate the volume of blood in the stain which is extracted in order to compare amounts per ml of blood.

II. LITERATURE REVIEW
Genetic markers currently employed in the forensic laboratory include the ABO, Rh and MN antigen-antibody systems, the hemoglobin and haptoglobin polymorphic protein systems, and the phosphoglucomutase (PGM) , adenylate kinase (AK) , erythrocytic acid phosphatase (EAP) , esterase D (EsD~, adenosine deaminase (ADA) , 6-phosphogluconate dehydrogenase (PGD) and glucose-6-phosphate dehydrogenase (G-6-PD) polymorphic enzyme systems (Culliford, 1971) . These systems were chosen for use in the examination of dried bloodstains because they are: 1) genetically controlled; 2) present and detectable in a bloodstain; and 3) supply individualizing information (Garner, 1979 (Schwinger, 1973, andWigmore, 1979) the method was found to be susceptible to erroneous identifications, caused by interfering fluorescence and the judgment of the observer. This method has been reported to be 40-80 % accurate in blind studies (Schwinger, 1973, andWigmore, 1979 The application of RIA techniques for steroids in forensic serology was ~eported independently by Brown (1976) and Stuver (1976) at the 28th annual meeting of The American Academy of Forensic Sciences. In the next year at the 29th annual meeting, Brown (1977) and Schehr (1977) independently announced the further application of steroid RIA techniques in bloodstains for the purpose of sex identification.
In the same year Moller (1977)  The clinical potential use of antibodies has been greatly expanded with the production of antibodies which are directed against small molecules of therapeutic and diagnostic interest, such as steroids, prostaglandins, thyroid hormones, cyclic nucleotides and certain drugs (Ekins, 1974).
It is the discriminating ability of antibodies to recognize one particular material as distinct from any other which is the basis for their use as an analytical tool in all immunologic methods. Although there is a limitation on an antibody's ability to discriminate between similar structures there are examples where extremely small alterations in the structure of a molecule will drastically change its binding to the antibody. Antibodies are not usually preformed in an individual but are produced in response to the entry of the foreign material into the body. Material that induces an immunoresponse is called an immunogen. The substance that is bound by antibodies is called the antigen. Most substances are both immunogen and antigen, but these two properties are distinct and not all antigens have the ability to initiate the formation of antibodies (Thorell, 1978).
Antibodies comprise a group of related proteins in blood plasma which are collectively referred to as immuno-  (Thorell and Larson, 1978).
The antibodies produced in response to an immunogenic stimulus constitute only a very small fraction of the total antibody population occurring in the plasma of an animal.
In most assays, however, there is no need to isolate a particular fraction from the other antibodies present, since the assay system only measures the binding of one particular radioactive antigen (Playfair, 1974). Other antibodies present will not interfere in this binding between the labeled antigen and its particular antibody. These mutually associated antigens and antibodies are called homologous antigen and antibody. In certain assay systems, one may utilize antibody that is raised against a different immunogenr in this case the antigen and antibody are called heterologous.

Radioimmunoassay (RIA) and related competitive protein-
binding methods began about 20 years ago as a cumbersome research methodology in two specialist radioisotope centers located in New York and London (Ekins, 1974). These methods represent a common analytical approach of great sensitivity that has been applied to the measurement of more than 200 substances, many of which cannot be assayed by other techniques. It is essentially impossible to exaggerate the diversity of applications for this methodology. Virtually every branch of medical and biological research has been affected by these techniques.
In particular, endocrinology has been greatly enriched by the new knowledge that has come as a direct result of RIA methods. These methodologies are being introduced into clinical medicine at a rapid rate, and the growing commercial availability of radioassay kits will revolutionize the routine practice of hospital and research laboratories (Challand, 1974).
The term radioimmunoassay was coined by Berson and Yalow (1960) to describe their original methodology. However, this term is . adequate for only a small portion of related assays which are being used today. There has been an effort to select a descriptive word or phrase that would be inclusive of all the various assay types, but it is now felt that the proper individual terms be used for describing a particular assay (Ekins, 1974 andThorell, 1978).
The term radioligand assay will be used interchangeably with RIA.
In order to establish a radioligand assay, three prerequisites must be met: (1) a receptor must be available that specifically binds the ligand (antigen) to be measured; (2) the ligand must be labeled with a radioactive nuclide (radioligand); and (3) separation must be achievable between the ligand bound to the receptor and the ligand that is unbound (Thorell, 1978). The RIA generally consists of a competitive type radioligand assay, although a noncompetitive type is sometimes employed.
The competitive assay is the most common type of procedure used in RIA. In this case, the . specific receptor is an antibody directed against the ligand (antigen) to be measured. The ligand is injected into a species of animal that sees this substance as foreign (iwmunogenic) and produces a specific antibody against the ligand. The radioligand is produced by coupling a radioactive nuclide to the ligand as a marker so that the ligand can be measured.
The basic principle of RIA is the competition between the radioactive and nonradioactive ligand for a fixed number of antibody sites. This interaction is represented schematically in Figure 1.
If increasing amounts of nonradioactive ligand and a fixed amount of radioactive ligand are allowed to react with a constant amount of antibody, a decreasing amount of radioactive ligand is bound to the antibody. This can be considered as a progressive dilution of radioligand bound to the receptor, or as a progressive dilution of radioactive material with nonradioactive material, thereby reducing the specific activity of the ligand bound to the receptor.
There is usually an incubation period for most assays, during which time the binding reaction between receptor anq ligand comes close to equilibrium. This incubation varies in time and temperature requiring anywhere from 30  (Parker, 1976). In general, the more dilute the receptor, the more sensitive the assay. At these low concentrations, longer incubation times are necessary in order to achieve a sufficient interaction between ligand and receptor.
After the binding reaction is completed, the separation of free from antibody bound antigen must be achieved in order to determine the partition of the radioactivity.
Various techniques have been employed for this purpose (Radcliffe, 1974). In many cases the antibody-antigen complex is precipitated out of solution by the addition of a saturated ammonium sulphate solution.
In some circumstances, a second incubation period is required, as in the double antibody techniques. When the antigen is a steroid, separation is achieved through selective adsorption of the free antigen onto dextran coated activated charcoal (Kushinsky, 1975). The amount of radioactivity of the bound fraction is then determined. There is an inverse correlation between the amount of radioactivity bound and the concentration of native ligand present in samples. This relationship is usually not linear. In order to be able to measure the amount of ligand present in an unknown biologic sample, such as plasma or urine, the assay must include a standard series of tubes that contain known amounts, in sequential increments, of the substance to be assayed (Bangham, 1974). Based on the radioactivity bound to the receptor in these standards, a standard curve is drawn and the amount of unlabeled antigen in a sample can then be determined by interpolation from this curve.
In the noncompetitive assay technique the antibody is usually labeled instead of the ligand. These methods are particularly useful when the ligand cannot be labeled easily, as is the case with most viral antigens (Thorell, 1978). The receptor in this method is coupled to an insoluble, solid phase matrix. An additional set of receptors is labeled with radioactivity.
In the initial phase the solid phase receptor is incubated with the sample containing the ligand to be measured until all ligand is bound to this receptor. The second step is to add the additional set of radio-labeled receptors which are soluble. This radioreceptor will adhere to any bound ligand in the preparation. The technique requires that the ligand be suf ficiently large enough to have two or more molecular regions that can act as combining sites (divalent or polyvalent).
After the second incubation the matrix is washed to remove all nonbound radioactivity. The remaining radioactivity bound to the matrix is measured. A standard curve is constructed by running a series of standards with known incremental amounts of the ligand. The standard curve for this assay is directly proportional to the concentration of ligand present; that is, the larger the amount of ligand present, the larger the amount of radioactivity bound to the matrix (Thorell, 1978).
RIA and related clinical methods derive their unique usefulness from their very great sensitivity in relation to other chemical techniques. This sensitivity is related to the extremely sensitive methods for the detection of radioactivity (Ekins, 1974). It is possible to measure the rad . t' · t 125 I · t ' t ' small as 10-4 g (or ioac ive iso ope in quan i ies as approximately l0-12 M) with good precision. Since the serum concentrations of most hormones is on the order of 10-9 to 10-lOM, detection of hormones at the lowest physiologic concentration is now possible. The quality of any assay is determined by four factors: sensitivity, specificity, precision, and accuracy. Specificity is predominantly a property which is dependent on the receptor. It is the ability of the assay to measure one specific compound and no other. The specificity or lack of it is a major problem in some assays. For many biologically important compounds, a number of compounds exist with similar chemical structures. Steroid pormone assays are a prime example where this problem occurs (James, 1974). There are many steroid derivatives with small differences in structure but with marked physiologic differences. In this situation there are two means of enabling the accurate measurement of a single steroid--the production and selection of highly specific antisera, which requires significant effort especially in the formation of a suitable immunogen. However, this cannot be accomplished to the point of exclusion of similar substances and therefore the situation is sometimes improved by chemical separation to isolate the compound of interest from the interfering substances. Solvent extraction, washing procedures and various types of chromatography are the procedures most often used (Thorell, 1974).
The sensitivity of an assay is limited by two factors: the affinity of the receptor and the specific activity of the radioligand. The affinity of a receptor reflects the energy in binding between the ligand and the receptor, the firmness of the binding. To be able to measure minute amounts of hormones and other biologic compounds with adequate sensitivity the energy of binding has to be very high.
This often is the limiting factor for the sensitivity of an RIA (Thorell, 1974).
The specific details involving the production of proper immunogens, specific antibodies, the preparation of the radioactive ligands and the kinetics of the antibody-antigen interaction is beyond the scope of this review. They are clearly detailed in several excellent reviews on RIA and related techniques (Ekins, 1974;Bangham, 1974;Hunter, 1974;Playfair, 1974;Parker, 1976;Dawson, 1978, andThorell, 1978 They not only produce labels of high specific activity but permit the use of simpler counting methods (Marks, 1974).
In pharmacology RIA offers a sensitive, precise and specific method for measuring concentrations of drugs and other significant substances in tissues and blood. This allows for research to be conducted into their distribution, metabolism and pharmacokinetics and, increasingly, for monitoring drug therapy (Marks, 1974 Although the major obstacle which researchers face in their attempt to develop RIA for proteins and peptides is in the isolation of the pure compound, the main problem confronting those developing RIAs for drugs and steroids is in the production of a suitable antisera. These compounds are rendered immunogenic by chemically coupling them to a larger I antigenic molecule. This usually involves the formation of a peptide bound between the drug or steroid and the protein, usually but not always, bovine serum albumin (BSA). The point of conjugation often has an important influence on the degree of specificity of the antibody produced; this is especially true with steroids. It has been only in the last five years that highly specific, low cross-reacting, antibodies have been produced utilizing immunogens conjugated at different points on the haptene molecule (Ekins, 1974). Another difficulty encountered when dealing with these compounds is that most cannot be radioiodinated directly. This difficulty is overcome by the use of 14 cor 3 tt-labelled drugs or steroids. Currently drugs or steroids are being coupled to compounds containing a p-hydroxyphenyl group that can be radioiodinated (Hunter, 1974).
As stated above a major problem with steroid RIA is the cross-reactivity of related steroids with the antibodies produced (James, 1974). For example in the RIA of testosterone the major cross-reacting steroid is dihydrotestosterone which initially showed a 100% cross-reactivity with antisera manufactured by New England Nuclear (NEN, 1975).
This cross-reactivity has recently been reduced to 56% through production of antisera to a different conjugate.
This problem is also evident in the RIA of the estrogens.
Here the three major estrogens: estradiol, estrone and estriol all cross-react significantly with antisera produced against an estradiol conjugate. When dealing with the measurement of these steroids, it is sometimes necessary, depending on the specificity required, to fractionate the sample extract by column chromatography prior to the assay. With increased antibody specificity the chromatographic step can be eliminated (Castro, 1973;Stahl, 1975 andSheldon, 1977) which is desirable when dealing with small samples.
Many of the RIAs developed for pharmacology also have a potential for use in the forensic sciences. Those RIAs developed for morphine and opiate alkaloids, lysergic acid derivatives, tetrahydrocannabinol (cannabis), barbituates and other drugs of abuse provide a screening method to detect the particular drug in a suspected abuser, in an overdose case or in blood samples found at a crime scene.
Blood or bloodstains are one of the most frequently encountered types of evidence at crime scenes; especially crimes of the mo~t serious nature such as homocides, assaults and rapes (Saferstein, 1977). The determination of other marker systems in these samples, such as sexual origin, will complement the evidence provided by the identification of the major blood factors. As stated previously the RIA method offers an accurate and sensitive method for sex determination based on the quantity of testosterone and estrogens present in a quantity of bloodstained material.
Because the steroid molecule is highly stable (Morrison, 1973) it is capable of withstanding the effects of drying in a bloodstain. The proper preparation and extraction of the stain with ether enables the investigator to measure each steroid fairly accurately. The results obtained are then calculated as a ratio for each sample and used as a basis for discriminating between male and female samples.
When working with any RIA method it is essential to know the limits of the antibody being employed, especially in reference to its cross-reactivity with compounds other than the one which is being measured.
In working with blood it is not uncommon to have several compounds present in the extract, and therefore in the test system, which are capable of cross-reacting with the antibody and causing an overestimation of the quantity of the compound being measured (James, 1974). This would also be true when working with bloodstains since diethylether will extract from a bloodstain many _nf the organic soluble steroids present.
The cross-reactivity of many of the naturally occurrin~ steroid hormones with the antisera used in this study has been reported by the manufacturer, New England Nuclear (protocols 1975, 1976). However, the question of crossreactivity of this antisera with synthetic steroid hormanes, used in oral contraceptives, has not been reported and is undertaken in this study.
The compounds ethinyl estradiol (EE 2 ), ethynodial diacetate, mestranol, norethindrone (NET), norgestrel (Nor-G) and norethylnodrel are the most commonly used synthetics in oral contraceptive preparations (Murad, 1975). The first and third are synthetic estrogens and the remaining four are synthetic progestins. Either group is capable of cross-reacting with the antiserum used in either the testosterone or estrogen assays (see Figures 2 and 2A) .
The literature available on these compounds is extensive and a good review of their metabolism has been published (Ranney, 1977). EE 2 , NET and Nor-Gare the most representative of this group since mestranol is converted to EE 2 and the others to NET in vivo. Reports concerning the plasma concentration of these drugs after administration is limited. This information is necessary in order to perform significant studies dealing with cross-reactivity. Verma (1975) reports the plasma levels to range from 144 to 248 pg/ml after 6 hours in women receiving 50ug of EE 2 or mestranol, through the use of a competitive protein binding radioass~y for EE 2 . RIA procedures developed for NET (Nygren, 19ii and St~nczyk, 1978) show plasma l~vels to rise and fall rapidly and attain peak levels of approximately 15ng/ml within 1.5 hours after ingestion of a lmg dose. Finally, Brenner (1977) reported levels of Nor-G, as determined by RIA, to range between 8-12ng/ml plasma in women receiving 500ug per day for twenty-one days. The normal daily dose for each of these compounds as reported by Murad (1975) are as follows: EE 2 : 20ug to lOOug with an average of SOug; NET: 500ug to 10,000ug with an average of l,OOOus; Nor-G: 75ug and 500ug.
In addition to their cross-reactivity potentials these compounds could affect the assay in another way. Briggs (1976) has reported that the administration of these synthetic hormones cause changes in the estradiol fluctuations seen in the female cycle.
In women being administered these preparations there are no estradiol peaks observed, due to LH suppression, although basal levels remain the same. These drugs affect progestrone levels in the same manner.
Testosterone levels have been observed to increase slightly.
In essence these drugs could cause an over or underestimation of the actual T and/or E levels present in a stain from a female who is taking such medication.

III. EXPERIMENTAL
During the course of the study blood samples were obtained from two groups of volunteers.   Table 2 Antiserum Titer Determination  (McArthur, 1970 andNEN, 1975).
The tritium labeled tracer was supplied in a benzeneethanol  Table 3, to provide a range of standards appr6priate for the assays.
The extraction of the artificially prepared bloodstains were carried out in labeled 13Xl00rnrn disposable test tubes. From the twelve 2X2cm squares for each individual one was extracted for the testosterone assay and three were extracted for th~~estrogen assay due to the nearly three-'t.f fold differences in plasma concentration (NEN, 1975). The remaining pieces were used in the age study assays. Prior to each extraction, O.lml of the appropriate recovery tracer was added to the pieces of cloth in the tubes and allowed to dry. The stains were then extracted with 2,2 and lml portions of an organic solvent. Initially, methylene chloride was used for the testosterone extraction and   2 The assay tracers were prepared at 4000 cpms for the open and single-blind studies and at 10 4 cpms for the cross-reactivity study. 3 Antiserum is added last in every case. All tubes were incubated at 4°C for approximately 24 hours. 4 standards B-J f{om Table 3 were used in all three studies. The concentrations of synthetic steroi 1 ' . <;! shown in Tables 5 and 6 were treated as standards in the crossreactivity study. 5 nried down aliquots of the ethanol reconstituted extracts. w \.;, steroids mentioned previously, was undertaken. The six synthetic steroids supplied to us, through the various pharmaceutical companies mentioned, were each weighed to the nearest O.Olmg. Each was dissolved in an appropriate volume of a benzene:ethanol (9:1) solution to obtain a concentration of lmg/ml. These stock solutions were appropriately labeled and serially diluted as outlined in Tables 5 and 6 to give ranges of concentrations which would be suitable for use in the crbss-reactivity studies.
The determination of the cross-reactivity of these synthetic steroids was accomplished in two separate assays with the testosterone and estradiol antiserums.
Each of the six steroids, in the concentration ranges shown in Table 5, were tested against the estradiol antiserum. The assay of these compounds were run concurrently with a set of estradiol standards prepared as shown in Table 3. In the crossreactivity studies with the testosterone antiserum, there was a need to conserve some of the components of the testosterone RIA kit and therefore only three representative synthetic steroid~~' EE 2 , NET and norgestrel were te~ted. The ranges of concentrations shown in Table 6 were assayed concurrently with a ~et of testosterone standards (Table 3).
The protocol for the cross-reactivity study is the same as outlined in Table 4 for the other two studies.           These adjusted values were put into ratio form (T:E) for the same sample at each age. The resulting ratios for all samples at each age group are shown in Tables 8 and 9 from the open and single-blind studies, respectively.
The T:E ratios from the single-blind study were used to determine the accuracy of this method to predict the sexual origin of the bloodstained samples in this study at each age level.
Since it was known that the numbers of male and female donors for this study were equal (eight), the ratios were independently divided into two equal groups at each age .level. The higher values were designated as male and the lower values as female. These designations were then compared with the correct sexual origin of each sample. It was found that the selections correctly designated within tha _2, 30 and 60 day age levels were ,;, # '%, 62.5% .

' "'
and 50% correct, respectively. A test for significance of a proportion was used to determine which of the percentages were significantly different from an expected 50% correct, since this was a single-blind study. The percent correct at the 2 day age level was the only value shown to be signifi-+ cant (z=2, critical z_ 05 = -1.96) within a 95% confidence leve 1.  Analysis of variance (ANOVA) , based on a two-way mixed design, was also performed on the mean T:E ratios (Table 10) from each study. The ANOVA analysis should demonstrate: 1. whether there was a significant difference between male . and female ratios, (conditions) , 2. whether there was a significant difference between ratios for the same samples at different age levels (trials) and 3. whether there was a significant interaction between the age of the sample and the distinction between male and female samples (trials x conditions) . The source tables for each ANOVA are contained in Appendix C (Tables 14 and 15).
The T-test analysis for differences between males and females showed that there was no significant difference between male and female ratios only at the 60 day age level in confidence level. _g.Jn the ,, T-test analysis for differences between age levels several comparisons shov-.Bd no significant difference at a 95% confi- In the open study there was no significant difference bet,:ween all age comparisons for female ratios.
In the blind study no significant difference was seen between the 2 and 30 day age levels for male ratios or between the 30 and 60 day levels for the female ratios. There was also a significant interaction between these two factors.
For the single-blind study, ANOVA showed significant differences (p=0.05) between males and females at each level and also between the same samples at different age levels; however, no significant interaction occurred between these two factors.
The results from the cross-reactivity study were handled in a slightly different manner than the results from the other two studies.
Instead of comparing the cpms of the standard and synthetic steroid samples to the total counts, they were compared to the zero standard counts to indirectly determine the percent of displacement of labeled antigen from the antiserum (Eq. II). The percentages obtained for each standard and synthetic steroid were then plotted against the corresponding log-dose of the steroid added (ng) on semi-logarithmic paper. The curves obtained in the testesterone cross-r~a~ivity for testosterone, ethinyl estradiol, norethindrone and norgestrel are displayed in Figure 14.
The curves obtained in the estradiol cross-reactivity study for estradiol, ethinyl estradiol, mestranol, norethindrone, norethynodrel, norgestrel and ethynodiol diacetate are displayed in Figure 15.
The percent of cross-reactivity any compound shows with an antiserum is determined at 50% displacement of the Only one of the synthetic steroids tested showed 50% displacement at the concentrations tested (Table 11).
Ethinyl estradiol showed a 5.00% cross-reactivity with the estradiol antiserum. The other compounds tested showed less than 0.01% cross-reactivity with either the testosterone or estradiol antisera. Values for these compounds were taken from the Protocols supplied with RIA kits. ** NS-Not significant at the therapeutic level or at higher levels tested (less than 0.01% cross-reactivity). The recovery results from each assay showed that steroids are easily recovered from bloodstains when extracted with diethylether. The average recovery for both testosterone and estrogen was about 70% or better (Table 7). In three assays the recoveries were less than 50%.
In the testosterone assay of the 2 day open study the recovery was only 46.5%. This extraction was performed using methylene chloride. Both testosterone and estrogen assays of the 30 day single-blind study showed less than 30% recovery for unknown reasons.
In the clinical use of these RIA kits, recoveries of less than 40% are considered too low and should be repeated; however, in this study this was no~ possible and the assays were carried out as usual.
Radioimmunoassays of the sample extracts produced data which indicated that significant amounts of each steroid are detectable. These amounts were adjusted accordingly and put into ratio form, T:E (Tables 8 and 9), for compari- In imrnunoassays the rule is that the more dilute the antiserum the more sensitive the assay. Therefore, when the dilution produced greater than 50% binding the assay was less sensitive than if the dilution showed 40% binding.
The cross-reactivity experiment showed that there is very little direct interference from synthetic steroids.
Only ethynyl estradiol showed any significant cross- id e ntify the individual from whom a bloodstain originat'ed.
With the development of specific antisera the identification of several genetic markers is easily accomplished. This principle can also be applied to physiological factors which vary as a result of genetic programming. To this end this initial investigation was devised to apply such techniques, through RIA methodology, for determining the sexual origin of a bloodstain based on the amounts of testosterone and estrogens detected.
The limitations of this particular study lie in the time factors involved. The effects of age in terms of the age of the bloodstain, the age of the person from whom the blood sample was obtained and the age of the RIA kits utilized, have been discussed. The first two are beyond the control of the investigator and the third must be carefully monitored in order to have consistence between investigations. It is recommended that future studies utilize larger numbers of subjects as blood sources and that the subjects be screened as to whether they are taking specific medication, which with the RIAS . In this way a true indication could be made as to the accuracy of this method in both the presence and absence of interfering substances . It is hoped that this study will contribute to making this p r ocedure as common in the forensic laboratory as are the procedures for identifying the components of the ABO system.

VI. SUMMARY AND CONCLUSIONS
An investigation -wasmade concerning the ability of commercially available RIA kits to quantify testosterone and estrogens in extracts of dried bloodstains, for the purpose of determining the sexual origin of those bloodstains.
Whole blood obtained from male and female volunteers, was used to prepare twelve bloodstains on pieces of 2 X 2 cm white cotton linen, for each blood sample. Bloodstains were extracted and RIA carried out on the extracts at 2, 30 and 60 days after their initial preparation. Resulting ·values for testosterone and estrogen from each extract were put into T:E ratios. Ratios were compared for male vs. female differences and also for differences between samples from the same blood source at the three (2, 30 and 60 days) age levels. The accuracy for discriminating between bloodstains from male and female sources and the direct interference of several synthetic steroids were also determined for this appli€~}ijion of the RIA technique.
A significant difference was seen between ratios from male and female sources in five of the six comparisons.
Only the final comparison, 60 day single-blind assay, showed no significant difference between male and female bloodstains. There was also a significant difference between ratios for bloodstains from the same source at the different time intervals (2, 30 and 60 days Since one-fourth of the reconstituted extract is counted. ** Corrected zdro standard count was used for standard and synthetic steroid counts in the cross-reactivity study. *** Since the dri e d extract is reconstituted to 2.0ml with absolute ethanol . •