Forensic Science International,

51 (1991) 1 - 12
Elsevier Scientific Publishers Ireland Ltd.

SEMEN FINGER PRINT

S.H. DOSS and N.A. LOUCA*

National Research Center, Dokki, Cairo (Egypt).
(Received August lOth, 1990)
(Accepted April 2Oth, 1991)

Summary
A correlation between testosterone (T) and dihydrotestosterone
(DHT) exists in seminal plasma.
These androgens play a role in sperm maturation. The TlDHT ratio is different from one person to
the other, due to the heterogenicity of seminal plasma which stems for the most part from the male
accessory sex glands, the prostate and seminal vesicles, and also from the epididymo-testicular
unit.This ratio is useful in identifying the persons semen. Consequently, the steroid values from assailant semen or accused persons and semen on the victims clothes are of cardinal importance in
matching. The results reported include data on the validation of this technique as a tool for semen
identification. Coital and masturbated semen were correlated, and consecutive coital semen were
also analysed. It may be concluded that the T/DHT value is specific for each person.
Key words: Semen androgens; Constant T/DHT ratio; Matching semen stains; Semen testosterone
and dihydrotestosterone
semen finger print

Introduction
Laboratory findings indicating that stains on a victims clothes are seminal
fluid, detected by ultraviolet light, flourescence test, microscopy for spermatozoa, acid phosphatase test etc. are not sufficient to identify the ejaculator.
Recently, the application of DNA finger-printing individualization of probands
has been a routine procedure in semen analysis in forensic hemogenetics [l],
especially when the pattern of restriction fragment length polymorphisms
(RFLP) detected, is identical to that observed with DNA prepared from blood
of the suspected male. Therefore, RFLP analysis can be used to exclude or determine probable identity of an assailant in rape cases [2].
Laser, single photon fluorimetry c+n differentiate between different semens.
It gives a superimposable band for the original semen with its diluted sample
with deionized water, or the extract of such dried stains with deionized water [3].
Presat address: Hospital Saint Andre, 1, rue Burguet 33075 Bordeaux, France.
0379-0738/91/$03.50
0 1991 Elsevier Scientific Publishers
Printed and Published in Ireland

Ireland Ltd.

Aim of the work

In the present communication we would like to report another tool for semen
finger printing through the hormones present in the semen. For our study we
chose two steroidal hormones, because these are more stable than the peptide
hormones: also they are not affected by temperature and atmospheric oxygen.
These two steroidal hormones are testosterone
and its precursor dihydrotestosterone. The TlDHT ratio is a constant value for each person in view of their
importance in spermatogenesis
and sperm maturation [4,5], which depend on
pituitary activity controlling the hormonal balance of the body. In addition they
are potential parameters of testicular hormone production.
The major source of DHT in seminal plasma, is the 5cr-reduction of testosterone by epididymal tissue [6]. Theoretically, a deficiency in testicular hormone
production and/or action would result in abnormalities in seminal plasma androgen levels or in the efficiency
of conversion
of testosterone
to
dihydrotestosterone
[7]. All these will lead to different TlDHT ratios from person to another.
Consequently, if we suppose that T is semi-constant in the seminal fluid of normospermic persons, DHT will be completely different for each person, hence the
ratio of TlDHT will give a fraction characteristic
for each semen. Thus
gonadotrophin response to gonadotrophin-releasing
hormone may be a better
measure of the integrity of the hypothalamic pituitary testicular axis 181. Hence
comparing this fraction obtained from assailant semen with that found on the
victims clothes, will give decisive proof of similarity.
The above suggestion is based on the fact that seminal plasma testosterone
derives mostly from the accessory glands, whereas about two-thirds of seminal
plasma DHT is issued from the epididymo-testicular
unit. This is clear from
T and DHT estimations in split semen specimens. DHT was higher in the first
fraction containing prostate and epididymal fluid than in the second fraction,
while this was not the case for testosterone. This was emphasized by the results
of La Lannou et al. [9], whereas DHT concentrations were significantly lower
in azoospermic and oligospermic samples, while T showed no significant differences.
The required differentiation
is often conducted between two persons. HOWever, in cases where the number of persons exceeds two, the maximum number
of potential matches will be limited to ten persons only. For fresh seminal fluids
it may be easily estimated through radioimmunoassay (RIA). Nevertheless, if the
semen is on sanitary napkins or under-wear, extraction of whatsoever, the I
amounts will be accompanied by the proportional DHT and we can achieve the
same percentage TlDHT as that taken from the fresh semen.
Actually, when these fractions are compared with the accused persons (usually
not exceeding ten persons) and the specific fraction obtained from the victims
under- wear, we can identify the assailant.
Materials and Methods
Our material consisted of 10 normal and 68 abnormal specimens. Normalcy
was defined as follows: > 40 x lo6 spermatozoa/ml semen, > 60% motility, >

70% vitality and 50% normal morphology. The group of abnormal specimens included azoospermic and infertile persons, and those with hydrocele and
varicocele.
Morning semen samples were obtained by masturbation, after 4 days sexual
abstinence, from 150 patients attending the fertility laboratory. Samples were
kept at 37C for 10 min to liquefy, then examined prior to centrifugation. The
supernatant seminal fluid was separated and was stored at - 25C after adding
a very small amount of sodium azide for preservation.
After thawing, testosterone and dihydrotestosterone were measured by the
method of La Lannou et al. [9], using Amersham testosterone and dihydrotestosterone radioimmunoassay kits.
Isolation and purification of the steroids in the ethereal layer (for 7 samples)
were carried out using the modified procedure of Parker et al. [ 111and Anderson
et al. [12] In the other samples, RIA was carried out without purification.
The T/DHT ratio did not show any marked changes before or after
chromatographic separations. Difference was within experimental error. For extraction of the dried semen stains on cotton or nylon underwear or sanitary
napkins, 2-cm samples were placed in screw capped tubes and soaked in 3 ml
deionized water for 3 h, then 5 ml of ethyl ether was added and centrifuged at
10C (3500 x) for 15 min.
Tubes were chilled in acetone-carbon dioxide snow to freeze the aqueous layer.
The ether layer was decanted and dried under a stream of nitrogen. Fresh ether
(3 ml) was added to the extraction then centrifugation and freezing of the
aqueous layer was repeated with the second ether extract combined with the
first ether extract. The combined ether layer was evaporated under a stream of
nitrogen.
The induced stains, for comparison, were made by adding blind samples
haphazardly choosen by another colleague, on cotton cloth, nylon cloth and
sanitary napkins. These were then left to dry at room atmosphere 25-28OC for
10 h, then kept in petri dishes.
Measurements and calculations were done with the Berthold-12-gamma
counter.
Results and Discussions

It has been reported [13] that there is no difference between testosterone

levels in young fertile men and sexually-disturbed middle-aged men (52-68
years). Testosterone concentrations showed no significant differences between
fertile and infertile semen, i.e. samples [9,14] from varicocele patients and controls [15]. However, Bruno et al. [16] have reported significantly lowered concentration of testosterone in severe oligozoospermia and in teratozoospermia.
The studies of Ragni et al. [17] have shown a decrease in semen testosterone
in men with inflammatory bowel disease and who have been treated for a long
time with sulfasalazine and who showed abnormal sperm motility and morphology with a predominance of large-headed spermatozoa. Medicaments like
cimetidine elevate testosterone levels during therapy [18] and persons with alcohol dependency syndrome [19] showed low testosterone values.

4
TABLE

NORMOSPERMIC

Sample No.

T (nglml)

DHT (nglml)

TIDHT

1
2

1.05
0.32

1.70
0.70

0.61
0.45

3
4
5
6

0.73
0.43
0.30
0.43

0.80
0.77
0.78
0.90

0.91
0.55
0.38
0.47

7
3
9
10

0.46
0.26
0.52
0.64

0.80
0.60
0.98
0.86

0.57
0.43
0.53
0.74

For our study, the quality of the coital semen should be similar to that of
masturbation origin, as shown by Purvis et al. [20].
Testosterone level in seminal plasma for each person is constant, as the assessment of semen quality using the conventional parameters of accessory gland
function can be made with ejaculates provided by masturbation or coitus [20].
Phenoxybenzamine when administered in doses up to 20 mglday, will cause

TABLE

INFERTILE
T hdml)

DHT (nglml)

TIDHT

1
2

10.68
0.35

0.88
0.86

0.77
0.40

3
4
5
6

3.25
1.02
0.54
1.14

0.92
0.85
0.79
0.89

3.53
1.2
0.68
1.28

7
8
9
10

0.12
0.20
0.72
0.44

11
12
13
14

0.24
0.24
1.05
0.44

15
16

0.43
0.15

0.72
0.84
0.99
0.74
0.89
0.70
1.26
0.78
0.90
0.70

0.16
0.23
0.72
0.59
0.26
0.34
0.83
0.56
0.47
0.21

Sample No.

aspermia following male ejaculation but will not affect the hormonal balance of
the body [19].
Samples from the victim will be due to coitus while the check samples for correlation will result from masturbation, but there will be no difference in T level
in the two samples obtained by the different routes.
It has been postulated that free T ranges from: 7.0 f 2.8 to 7.9 f 2.3 ng/
100 ml.
However, seminal plasma testosterone for Egyptians, whether for normal or
diseased persons (cf. Tables l-4), shows in general higher values than that
reported for Europeans by Purvis et a1.[6,9], as compared with Salem et al. for
plasma T and DHT, in Egyptians [lo].
This was clear in our study on testosterone levels in Iraqi people, supported
by the results of Abdel Aziz et al. [21].
It is clear from Table 5 that the amount of T in varicocele patients is greater
than DHT, which confirms the results of Hudson et al. [7] as men with
varicoceles have a deficiency in epididymal 5a-reduction of testosterone.
Regarding the second steroidal androgene DHT, its concentration in semen
depends on many factors. In azoospermic and oligozoospermic men it was
significantly lower than in normospermic persons [9].
Most of the seminal plasma DHT arises from the epididymal Ficr-reduction of
testosterone and a small portion arises from testicular production and accessory

TABLE

HYDROCELE
Sample No.
1

2
3
4
5
6
7
8
9

T (nslml)

QHT (ngfml)

TIDHT

0.41
0.46

0.88
0.78
0.82
0.98
0.87
0.99
0.90

0.46
0.58

0.45
0.75
0.44
0.86

10

0.44
1.48
0.26
0.86

11
12
13
14

0.62
0.69
0.44
0.40

15
16
17
18

0.93
1.23
0.59
0.14

19
20

0.20
0.29

1.90
0.80
0.98
0.92
0.94
0.86
0.87
0.98
1.46
0.98
.0.90
0.70
0.69

0.54
0.76
0.50
0.86
0.48
0.77
0.32
0.87
0.67
0.73
0.51
0.45
0.94
0.84
0.60
0.15
0.17
0.42

6
TABLE

AZOOSPERMIC
Sample No.

T (dml)

DHT (nglml)

TIDHT

1
2

0.67
0.77

0.12
0.22

5.58
3.5

3
4
5
6

0.68
0.68
0.38
0.22

0.20
0.14
0.13
0.12

7
8
9
10

0.12
0.53
0.26
0.35

11
12
13

0.24
0.22
0.15

0.09
0.24
0.12
0.14
0.17
0.19
0.12

3.4
4.8
2.9
1.8
0.16
2.2
2.1
2.5
1.4
1.1
1.2

gland secretion [8]. As was mentioned by Hudson [7], the TlDHT ratio should
give a rough index of the epididymal junction in men with a structurally intact
genital tract. This ratio is a more critical measure of the integrity of the
testicular epididymal unit than is a measure of the concentration of either steroid
[7], hence it will be useful as a measure of the ratio of TlDHT present in a dry
semen stain than in a fresh stain.
Tindall et al. [22] have proved that the epididymis converts T to DHT, hence
DHT is produced within the epididymal tissue and will be a characteristic value
for each person, as the denominator is different in each case. The ratio TlDHT
will differ from one semen to another.

TABLE

VARICOCELE
Sample No.

T (nglml)

DHT (nglml)

TIDHT

1
2

1.23
2.27

1.08
1.60

1.13
1.41

3
4
5
6

0.76
0.24
0.45
0.44

7
8
9

1.70
1.05
1.07

0.70
0.20
0.42
0.40
0.89
0.96
1.02

1.08
1.2
1.07
1.1
1.91
0.96
1.04

7
TABLE

AFTER

PURVIS

ET AL. 1301.

Testicular Tissue (nglg)

Epididymal

tissue (nglg)

DHT

TIDHT

437
1216

5.0
10.0

87
122

52.1
21.8

20.2
5.4

2.6
4.0

4.8

195
-

5.6
8.6

75
65

71.7
1.2
18.9
16.8

13.4
0.9
5.2
5.6

5.4
1.3
3.6
3.0

7.8
16.0
8.8
3.7

171
149
92
87

92.0
66.0
96.8
70.0

17.9
28.5
22.4
13.8

5.2
2.3
4.3
5.1

90
106

40.4
46.1
26.3
79.4

16.6
127.0
8.1
19.5

2.4
3.6

988
33
416
560

1332
2378
811
322
623
754
4901175

TABLE

6.9
7.1
5.29.6

DHT

TIDHT

T AND DHT LEVELS

IN FIVE RANDOM

SAMPLES

T
(nmoU1)

Sample

DHT

TIDHT

(nmoU1)

I
II

4.3

4.2

1.02

3.6

3.8

0.92

IV
VI
VII

3.2
3.6
1.6

2.9
3.3
0.8

1.10
1.09
2.00

TABLE
SEMEN

8
SAMPLES

FROM ONE VOLUNTEER

Semen sample

T (pglm0

DHT (pglml)

TIDHT

1st day
2nd day
3rd day
4th day

196
177
168
159

417
377
357
338

0.470
0.469
0.470
0.470

8
TABLE

SEVEN

RANDOMLY

Sample No.

CHOSEN DRIED AND FRESH

SEMEN

STAINS

Fresh semen

Dried semen stains

T WmU

DHT
@glmG)

TIDHT

T (nglml)

DHT
@g/ml)

TIDHT

0.40

0.53

0.75

0.64

0.86

0.74

II

0.32

0.66

0.48

0.43

0.90

0.47

III
IV
V
VI
VII

0.61
0.52
0.34
0.80
0.33

0.66
0.67
0.60
0.84
0.61

0.92
0.77
0.56
0.95
0.54

0.73
0.68
0.44
0.93
0.44

0.80
0.88
0.78
0.98
0.82

0.91
0.77
0.56
0.94
0.53

This view is supported by the data presented in Table 6 (after Purvis et al.
[20]). It is also supported by the results of Gastaneda-Pena et al. [23] where they
have postulated that TlDHT semen ratio = 0.42 and the correlation T = 0.55,
P < 0.01.

As a control for our measurements, T and DHT in 5 random samples were

measured (Table 7) by Dr. Oakey. The ratio of TlDHT was different in each
sample.

4.5
4.0
3.5
3.0
2.5
g
.
I-

2.0
1.5
1a
0.5
0.0

NORMAL
Fig. 1.

INFERWE

HYDROCELE

VARCOCELE

AmsPERMlA

Although the ratio of T/DHT was different

these results is suspect because:
1.
2.
3.

for each sample, the validity of

The degree of efficiency of testosterone extraction from the samples is

unknown.
The recovery measurements showed that DHT was extracted only poorly
from the samples.
There were indications, during DHT assay (which involves an oxidation
step with KMn04) that contaminants in the sample might have interfered
in the oxidation step (Oakey, R.E., private communication).

Coital semen samples from one volunteer gave the same ratio T/DHT as shown
in Table 8. Such results prove that semen androgen estimation is more reliable
than plasma androgen estimation, as was mentioned by Fox et al. [24].
The seven dried semen stains we chose randomly, gave the same ratio as their
original fresh semen, as shown in Table 9. Within these 7 samples, it is clear that
no ratio for T/DHT is similar, i.e. the ratio is specific for each person.
To make even more sure that the semen finger print is specific for different
persons, another study for the determination of TlDHT ratio in different semen
samples by gas chromatography-mass
spectrography is under investigation.

9 = .995x

- .002,

R-squared:

1,

f
i

c
t
./'

.i5.
1'

Fig. 2. The correlation coefficient (r)

this example the correlation between
was 0.9 indicating a very significant
correlation between the dried semen
for the ratio TIDHT, which strongly

indicates the strength of correlation between two readings. In

the dried and fresh values was done. For the testosterone it
correlation, while for DHT it was only 0.5 indicating weaker
values and fresh values. The exact correlation of 1 was found
supports the idea of a unique stable value for each individual.

10

Hi&i:

-5
*a
rl~rI~~fO
:
159

Sid.cev.:
15.811

196

Std.Drv.:

S .xi

33.817

Minimum :

t-laximun .
417

Mean:

Std.Dw/.:

.47
Minimum:

Maximum
:

,469

7.96s

rliU(UTlum: Riiw:

I-Ian:

I338

x1
Std.Erw:

4844E-4

.471

37

:~r~pg/IaL
'&fiir~i:
250
SW:
700

X2:DHT&rL
SrQ.Error: vvianc4:
16.908
Rarqt:
79

1143.583

sum:
1489

Cccf.Vx

..

5.035
S~~Warrd:
123250

cod.vu.:
Isa84

Cwlt :
4
??t'h$su~:

cow:
4

sumsquared:tliselg:
557711

X3 : T/DHT
StQ.Errbr: VW&C@:

Crf.Vu.:

2.422E4
Range:

SunlSquJrtiJ:Mi5sing:

.ODl

2.346E7

=Qm:
1.881

,103

.a884

Count:
4

1
I

Fig. 3. Coefficient of variation shows the extent of heterogenicity or variations among the different
readings = (SD/Mean) x 100. For the testosterone and DHT it is about 9%, while for the T/DMT
ratio it is (0.1%) which is very low indicating the consistency of readings for repeated tests.

Conclusion

The ratio of TlDHT in semen is specific for every male. The data present in
this communication are supported by the data reported by Purvis et al. (see Table
2 in ref. 25) for testicular tissue and epididymal tissue in different subjects, by
the results of Ando et al. (see Tables 1 and 3 in ref. 26), by the results of La Lannou (see Table 1 in ref. 9) and finally by the results of Bain et al. (see Figs. 1
and 2 in ref. 27).
It has been suggested that the normal steroid levels in seminal plasma
established by Diczflausy et al. [28], could serve as reference values in the hormonal evaluation of male infertility.
Statistical

analysis

Statistical analysis was determined by the paired t-test and Students t-test

Km.
Acknowledgement

The authors thank Prof. Dr. G. Sturm (University of Marburg, Germany) and

Dr. R.E. Oakey, (University of Leeds, U.K.) for checking some of the results and
their helpful discussions and the British Council for a research grant (Doss).
We are also greatly indebted to Professor Dr. Cyril H. Wecht for following the
maturity of this piece of research for more than 5 years.
References
J. Henke and L. Henke, DNA Fingerprinting Anwendung der Genetechnik in der forensischen
Hamogenetik, Biotech-Forum, 5, (1988) 115.
2 A. Giusti M. Baird, S. Pasquale, I. Balazs and J. Glassberg, Application of deoxyribonucleic acid
(DNA) polymorphisms to the analysis of DNA recovered from sperm. J. Forensic Sci. 31, (1986)
409.
3 S.H. Doss and N. Anis, Single photon fluorometry as a tool in semen finger print, International
Academy of Legal Medicine and Social Medicine, 50th Anniversary XIV II, 1988. Liege, Conference Abstracts, (1989) 146-147, Keyword, 36.4. Acta Med. Legales, (in press)
4 M.R.N. Prasad, N.Dinaker, R.Dinaker and F. Pajalakashmi, effects of androgens on the
epididymis. In: V.H.T. James (ed.), Endocrinology, Excerpta Medica, Amesterdam, 1977, pp.
417.
5 M.R. Rivarola, E.J. Podesta, H.E. Chemes and R.S. Calandra, Androgen metabolism and concentration in the semineferous tubules at different stages of development. J. Steroid Biochem.,
6 (1975) 365.
6 K. Purvis, B.M. Landgren, Z.Cehan and E-Diczfalusy, Indices of gonadal function in the human
male II. Seminal plasma levels of steroids in normal and pathological conditions. Clin. Endocrinol., 4, (1975) 247.
7 R.W. Hudson, K.A. Hayes, V.A. Crawford and D.E. MC Key, Seminal testosterone
and
dihydrotestosterone
in men with varicocele. 1nt.J. Androl., 6 (1983) 135.
8 J. Cohen, D.Delafontaine and J. Grenier, Artificial insemination with split ejaculate: a clinical,
sperm and hormonal story. Int. J. Androl., Suppl. 1 (1978) 165.
9 D. La Lannou, C. Massart, Y. Chambon, M. Nicol and H. Allannic, Testosterone concentrations
in human seminal plasma. Znt. J. Androl., Suppl. 3 (1980) 502.
10 S.I. Salem, O.M. Galal, T.J. Cole, S.M. Hussini, A.S. Hafez and A. Massoud, Response of plasma
testosterone and dihydrotestosterone
to some environmental factors. Arab J. Lab. Med., 13,
1

(1987) 95.

11
12
13
14

15
16
17

18
19
20

C.R. Parker Jr., J.O. Ellegood and V.B. Mahesh, Methods for multiple steroid radioimmunoassay. J. Steroid Biochmz., 6 (1975) 1.
CD. Anderson, B.R. Hopper, B.L. Lasley and S.S.C. Yen, A simple method for the assay of
eight steroids in small volumes of plasma. Steroids, 28 (1976) 179.
Z.T. Homonnai, N. Fainman, M.P. David and G.F. Paz, Semen quality and sex hormone pattern
of 29 middle aged men. Andrologia, 14 (1982) 146.
M. Pazzagli, G.G. Giusti, G. Forti, G. Fiorelli and F. Menchini-Fabris, Seminal plasma levels of
testosterone and 5-alpha.dihydrotestosterone
in azoospermic patients. Clin. Endocrinol., 11
(1979) 11.
S. Ando, G. Glacchetto, E. Beraldi, M.L. Panno et al., Testosterone and dihydrotestosterone
seminal plasma levels in varicocele patients. Acta Eur. Fertil., 13 (1982) 113.
B. Bruno, S. Francavilla, G. Properri, M. Martini and A. Fabbrini, Hormonal and seminal
parameters in infertile men. Andrologia, 18 (1986) 595.
G. Ragni, G. Bianchi Porro, M. Ruspa, G. Barattini, C. Lombardi and M. Petrillo, Abnormal
semen quality and low serum testosterone in men with inflammatory bowel disease treated for
long time with sulfasalazine. Andrologia, 16 (1984) 162.
C. Wang, C.L. Lai, K.C. Lam and K.K. Yeung, Effect of cimetidine on gonadal function in man.
Br. J. Clin. Pharmacol., 13 (1982) 791.
Z.T. Homonnai, M. Shilon and G.F. Paz, Phenoxy benzamine an effective male contraceptive
pill. Contraception, 29 (1984) 479.
K. Purvis, 0. Magnus, L. Morkas, T. Abyholm and H. Rui, Ejaculate composition after masturbation in the human male. Znt. J. Androl., 9 (1986) 401.

12
21

22
23
24
25
26
27
28
29

S.H. Doss, unpublished results cf. also, M.T. Abdel-Aziz, G.A. Tawadrous, S. Mostafa, M.S.
Karaa, W.M. El-Harizi, A. Abdel Aziz, S.M. El Haggar and I.A. Khair Serum and semen prolactin, testosterone, estradiol and zinc in impotent men. Bull. Egypt. Soc. Physiol. Sci., 5 (1985)
149.
D.J. Tindall, F.S. French and S.N. Nayfeh, Androgen uptake binding in rat epididymal nuclei,
in vivo. Biocha. Biophys. Res. Commun., 49 (1972) 1391.
G. Gastaneda-Pena, R. Alonso-uriarte Bedolla-Tovar and VCortes-Gallegos,
Blood and semen
androgen values in a young unmarried population. Arch. Invest. Med. (Mex.), 10 (1979) 215.
C.A. Fox, A.A.A. Ismail, D.N. Love, K.E. Kirkham and J.A. Loraine, Studies on the relationship between plasma testosterone and human sexual activity. J. Endocrinol., 52 (1972) 51.
K. Purvis, R. Calandra, S. Sander and V. Hansson, Androgen binding proteins and androgen
levels in human testis and epididymis. ht. J. An&ok, 1 (1978) 531.
S. Ando, C. Giacchetto, E. Beraldi, M.L. Panno A. Carpino et al., Testosterone
and
dihydrotestosterone
seminal plasma levels in varicocele patients. Andrologia, 15 (1983) 374.
J. Bain, M. Duthie and J. Keene, relationship of seminal plasma testosterone and dehydrotesterone to sperm count and motility in man. Arch. Androl., 2 (1979) 35.
W. Ying, M. Hedman. B. de la Torre, F. Jensen, P.H. Pedersen and E. Diczfalusy, Effect of
vasectomy on the steroid profile of human seminal plasma. Int. J. Androl., 8 (1983) 116.
Fiona Broughton Pipkin, Medical Statistics Made Easy, Churchill Livingstone, Edinburgh, London, 1984.