Last update: September 2, 2011 02:54:47 PM E-mail Print





E M van Tonder

Regional Veterinary Laboratory

Private Bag X528

Grootfontein College of Agriculture






Since its first isolation in South Africa (Van Rensburg, Van Heerden, Le Roux & Snijders, 1958), Brucella ovis has been regarded as the most important, if not the only cause of infectious epididymitis in this country. Vaccination with the attenuated Elberg Rev 1 strain of Brucella melitensis (Herzberg, Elberg & Meyer, 1953; Elberg & Faunce, 1957) introduced by Van Drimmelen (1960) proved to be a highly effective method of controlling the infection. Thus Van Heerden & Van Rensburg (1962) concluded that Elberg Rev 1 vaccine confers a solid immunity lasting for at least 2 years while a subsequent report (Van Heerden, 1964) claims the duration of immunity to be at least 4 years. Van Heerden & Van Rensburg (1962) and Van Heerden (1964) were convinced that B. ovis infection could be eliminated from infected flocks by meticulous vaccination.

The less favourable results reported some time later (Worthington, Van Tonder & Mulders, 1972) were attributed to faulty vaccination procedures, as was subsequently also established (Van Tonder, Worthington, & Mulders, 1972). The efficiency of Rev 1 vaccine in controlling B. ovis infection was again confirmed a few years later, when the results of regular investigations on stud farms where severe problems were encountered previously and vaccination has since been conscientiously practised, revealed the absence of B. ovis infection for at least 10 years (Van Tonder, 1977a, 1977b).

Renewed interest in the problem of infectious infertility, at that stage commonly regarded as B. ovis infection, started when the author was employed as state veterinarian at De Aar and testing of rams for genital soundness was included as part of a comprehensive disease control programme.

During these investigations it was found that on some stud farms clinical epididymitis reappeared while on many other farms it persisted, notwithstanding the introduction of Rev 1 vaccine and regular elimination of clinically identifiable cases.

Initially, the persistent incidence of clinical and sub clinical epididymitis was ascribed to the precarious application of Rev 1 vaccine. Failure to demonstrate B. ovis infection in the semen of these rams or homologous antibodies in their sera was looked upon as a reflection of the peculiar inconsistencies associated with the pathogenesis of an infection of this nature.

The gradual and often also sudden increase in incidence of genital lesions, and particularly a dramatic acute form in weaner ram lambs in certain studs, nevertheless caused a great deal of concern. This was particularly so because these young rams had had no contact with other sheep but their mothers in the preweaning period, and studs were often involved where B. ovis infection had never been encountered before or was successfully eliminated at some stage or other beforehand.

Repeated unsuccessful attempts to demonstrate B. ovis infection by microscopical examination of the semen smears of these rams which had been correctly immunized beforehand seemed to exclude the possibility of faulty vaccination or a breakdown in immunity. At the same time it was also established that B. ovis infection was not even involved in a high percentage of rams, particularly young weaner rams with clinical and sub clinical epididymitis which had not been vaccinated at all.

These observations, together with the presence of Gram-negative, non-acidfast bacilli (hitherto regarded as secondary invaders) in close association with neutrophils in the semen smears prepared from the majority of these cases led to the suspicion that other organisms might be involved.



The establishment of the Regional Veterinary Laboratory at Middelburg, Cape, at the end of 1966 created the opportunity to conduct comprehensive studies on a wide range of small stock disease problems on a local basis. Infectious infertility thus became the first condition to be fully researched.


At first a systematic preliminary investigation was initiated whereby all submitted cases of clinical genital infection were subjected to a full range of examinations and tests. In comparing the results of suitably vaccinated and non-vaccinated cases it became evident that B. ovis infection could not be confirmed in any of the vaccinated rams by anyone of the laboratory methods used, while it could be done with relative ease in at least a certain percentage of the non-vaccinated animals. Furthermore, it was found that a certain type of Gram-negative, non-acid fast, pleo-morphic bacillus was almost consistently present in the semen smears and cultures prepared from the semen of the majority of these cases, and particularly in those instances where B. ovis infection could not be established. On account of these findings the involvement of another bacterial agent, presumably the yet unidentified Gram-negative pleo-morphic organism, became strongly suspected (Van Tonder & Bolton, 1967).

In pursuance of these observations, a specific investigation was conducted on some stud farms where the problem of acute epididymitis persisted and even increased, in spite of meticulous Rev 1 vaccination. Abundant and often pure growth of a bacterial organism, closely resembling the organism isolated during the preliminary investigation, was obtained on cultures prepared from the microscopically infected semen of 21 clinically affected rams from 7 properties. Comprehensive bacteriological tests were then conducted on all these isolates, which not only proved them to be identical but also to closely resemble Actinobacillus seminis which was first described as a cause of ovine epididymitis in Australia (Baynes & Simmons, 1960). In the meantime a similar organism was isolated from the semen of a Dorper ram originating from the Orange Free State (Worthington & Bosman, 1968).

Comparative transmission experiments and detailed laboratory examinations confirmed the South African and Australian isolates to be identical and produced the final evidence on which A. seminis as a specific cause of epididymitis or infectious infertility in rams in South Africa was reported for the first time (Van Tonder & Bolton, 1968; Worthington & Bosman, 1968).

In order to determine the relative contribution of A. seminis and B. ovis towards the still prevailing problem of epididymitis and to ascertain whether some other bacterial organisms were also involved, two separate investigations were conducted. All the rams submitted to the Regional Veterinary Laboratory for examination for genital soundness during the course of 1967, in addition to the clinical and semen-smear examinations, were also subjected to a full bacteriological examination of their semen and to serological tests. In another investigation, all the young rams intended for sale during that year on four stud farms where sale rams were tested annually and where Rev 1 vaccine was consistently used were also subjected to the full range of tests.

Results of the examinations carried out on rams submitted to the laboratory showed that out of a total of 647 rams, 6,5 per cent had clinical lesions compatible with infectious infertility, 0,9 per cent had non-specific lesions and 15,1 per cent showed sub clinical infection (i.e. the presence of pus cells) in their semen. On bacteriological cultures of the semen, Actinobacillus seminis and Actinobacillus-like organisms (Ekdahl, Money & Martin, 1968) were isolated from 62,1 per cent of the rams with clinical lesions and 75,8 per cent of sub clinically infected rams, but could not be cultivated from the semen samples of rams devoid of microscopical evidence of infection (Van Tonder, 1977b, 1979a). On the other hand, Brucella ovis could only be isolated from the semen of 2,4 per cent of rams with clinical lesions and 4,1 per cent of sub clinically infected rams. No significant growth of any other bacterial organisms, other than obvious contaminants, was obtained on any of these semen cultures.

A total of 327 young sale rams were examined on the 4 selected stud farms. The incidence of clinical lesions in these rams varied from 1,9 to 4,2 per cent between the different properties and the rate of sub clinical infection from 13,0 to 34,7 per cent. Actino-bacillus seminis was isolated from semen cultures of 83,3 per cent of the clinically affected rams and 67,7 per cent of sub clinically infected rams but could not be obtained from the semen of those rams which showed no clinical and microscopical evidence of genital infection. No evidence of B. ovis infection could be found in any of these rams, which seems to confirm either that the infection never existed on these properties or that the young rams were fully protected by the vaccine (Van Tonder, 1977, 1979a).

Although these investigations were regionally confined and somewhat limited in extent, they nevertheless confirmed the significance of A. seminis in the aetiology of infectious ovine epididymitis and its major contribution towards both the clinical and sub clinical forms of the condition encountered amongst the rams examined.

In order to obtain more information on the importance of epididymitis as a problem in the sheep industry and the relative involvement of A. seminis and B. ovis as aetiological agents, extensive surveys and investigations were carried out, from which it was hoped to gain a fairly accurate assessment of the general incidence and geographical spread of these infections.

A specific clinical and serological survey was conducted on 409 randomly selected Merino, Dorper and Karakul stud farms distributed over 29 districts of the Cape Province and the Orange Free State. All these farms were visited during a predetermined period and all the rams present on the farm, whether breeding rams, sale rams or flock rams, were clinically examined and blood collected from them for serological tests. Since most farmers, particularly stud farmers, were familiar with the clinical form of epididymitis and, as part of the B. ovis control programme, were advised to dispose of such cases immediately, it was realized that only very recently developed and less conspicuous cases would be found on at least some of these farms and that the overall incidence revealed by the survey would therefore be a very conservative one.

A total of 34 501 rams were examined during the course of this survey and 2,5 per cent of them were found to have clinical lesions. On the serological test 27,1 per cent of the clinically positive rams reacted positively to A. seminis and 15,5 per cent to B. avis, while 15,8 and 6,6 per cent of the clinically unaffected rams were positive on the complement-fixation test in respect of A. seminis and B. ovis respectively (Van Tonder, 1973, 1977b, 1979b). Thus only 42,6 percent of the clinically affected rams could be accounted for serologically, being due to either A. seminis or B. ovis infection. The fairly large percentage (57,4 per cent) of rams with clinical abnormalities which failed to react to the serological tests for A. seminis and B. ovis was attributed to the presence of lesions of a non-specific nature, recently developed lesions i.e. where serological conversion had not taken place yet, long standing cases i.e. where homologous antibodies had already disappeared from the blood or lesions which had been caused by other types or strains of organisms.

Whatever the case might have been, the survey nevertheless indicated an economically significant general incidence of genital lesions, which, for the reasons explained, represented a very conservative reflection of the real situation. Apart from the fact that B. ovis infection both in clinically affected and clinically normal rams was still a problem to be reckoned with despite the availability of a highly efficient vaccine, these results also indicated that A. seminis was the more prominent infection involved.

The incidence of clinical lesions was considerably higher m- the flock rams of all breeds kept on these stud farms which was also proof of the fact that on these properties stud rams were more regularly examined and those showing clinical abnormalities disposed of as soon as possible (Van Tonder, 1973, 1977b).

Clinical lesions were found in rams on 60,1 per cent of the 409 properties included in the survey, while the serological tests indicated that exposure to A. seminis infection occurred on 74,8 and to B. ovis infection on 51,1 per cent of these farms. Since these properties were distributed over an area comprising 27 magisterial districts, the widespread occurrence of these infections became quite evident (Van Tonder, 1973, 1977b).

In a separate investigation, all rams submitted to the laboratory over a period of five years were subjected to as full a range of tests as was possible at the time.

Although these rams also included animals specifically presented with suspected genital problems, they were mostly young rams submitted for certification for sale purposes.

During the period in question, a total of 2 320 rams were subjected to clinical and semen smear examinations. On clinical examination 526 rams (22,7%) were found to have palpable lesions while a further 587 rams (25,3%) which had no lesions showed microscopical evidence of infection in their semen (sub clinically infected) and 1 207 rams (52%) were free of any sign of infection on both these examinations (Van Tonder, 1977b).

From the clinically affected rams of which the semen was examined bacteriologically, 61,8 per cent yielded cultures which were identified as A. seminis and 6,2 per cent had positive B. ovis cultures. In the sub clinically infected group of rams, 59,3 per cent of the semen cultures were positive for A. seminis and 1,9 per cent for B. ovis. Whereas all the clinically and sub clinically unaffected rams were negative for B. ovis on bacteriological cultures of their semen, A. seminis was nevertheless isolated from 13.8 per cent of these cases, although on microscopical examination they showed no evidence of active infection. These cases which mainly occurred in young rams and at the time could not be readily explained represented what subsequently, when the mode of transmission of A. seminis infection became clear, was described as a latent form of infection (Van Tonder, 1977b ).

During the first part of this investigation, the routine method of the complement-fixation test was employed to determine the presence of homologous antibodies, using either the standard B. ovis antigen or an antigen prepared from the original isolate of A. seminis. A total of 190 rams with clinical lesions were serologically tested in this way and 36,8 per cent reacted positively to A. seminis and 27,9 per cent to B. ovis. Likewise 30,0 per cent of the 80 sub clinically positive rams were positive to A. seminis infection and 8,8 per cent to B. ovis. Thus in all, and taking into consideration those cases which reacted positively to both tests, only 57,4 per cent of the clinical and 35 per cent of the sub clinical cases of epididymitis could be accounted for serologically (Van Tonder, 1977b).

As during the course of this investigation, comparative bacteriological studies indicated the existence of different strains of A. seminis, antigens prepared from six different strains of A. seminis were used in all subsequent serological tests (Van Tonder, 1973, 1977b). Since the incorporation of the different A. seminis antigens in the performance of the complement-fixation test, 263 rams with clinical lesions were tested serologically, and 63,1 per cent found to be positive to A. seminis infection and 24,7 per cent to B. ovis. Thus only 23,2 per cent of these rams were negative to both infections, which could have been due to non-specific lesions or the absence of antibodies due to too recently developed or too long standing lesions. Similarly the percentage of sub clinical cases which were positive to A. seminis infection increased by almost 19 per cent and at least 69,5 per cent could be attributed to either A. seminis or B. ovis infection (Van Tonder, 1977b, 1979b).

The wide-spread occurrence of both clinical and sub clinical epididymitis due to B. ovis and particularly A. seminis infection was again emphasized by the fact that the rams tested at the laboratory emanated from farms situated in 48 districts in the Cape and Orange Free State Provinces.

Both these studies confirmed that infectious epididymitis was still a major problem in the sheep industry. Evidence was also obtained to the effect that B. ovis still presented a problem and then mainly on properties where Rev 1 vaccination was inconsistently practised. It became quite clear, however, that in all events A. seminis infection played the most prominent role in the aetiology of ovine epididymitis in this country.

This was again substantiated by the results of continued investigations on 11 stud farms where severe problems of acute epididymitis were encountered despite the regular use of Rev 1 vaccine and the elimination of clinically affected cases. All the young sale rams on these properties were examined every year over a period of 4 years. The overall average incidence of clinical lesions in the 2 515 young rams thus examined was 6,9 per cent and the average incidence of sub clinical infection 33,8 per cent. Between the different farms and between years, the incidence of clinical lesions varied from 1,5 to 20,0 per cent and the incidence of sub clinical infection from 12,8 to 63,4 per cent. Actinobacillus seminis was obtained from almost all semen cultures from clinical and sub clinical cases which were examined bacteriologically, while B. ovis could not be isolated from any of these rams (Van Tonder, 1977b, 1979b).

On account of the fact that during the preliminary investigations bacterial organisms, which differed to some extent in their growth characteristics from the original Australian and South African isolates, were isolated in pure or almost pure growth from the infected semen of clinically and sub clinically affected rams within the same group and during the same outbreak, it was decided to examine the growth requirements of a large number of isolates in order to define a more specific set of growth and bacteriological characteristics for A. seminis. This was also deemed necessary as a large percentage of rams with clinical lesions, harbouring these somewhat divergent types of organisms, nevertheless reacted in the serological test to antigen prepared from the original isolate of A. seminis.

Baynes and Simmons (1960) described certain morphological, cultural and biochemical properties of the first strain of A. seminis isolated in Australia. In general, these characteristics were confirmed in subsequent reports (Livingston & Hardy, 1964; Worthington & Bosman, 1968; Van Tonder & Bolton, 1968). However minor differences Were recorded. Thus Worthington & Bosman (1968) found that their strain of A. seminis produced a clot but no acid in litmus milk, while the Australian strain (Baynes & Simmons, 1960) produced no such changes. Working with various original cultures isolated in South Africa, Van Tonder & Bolton (1968) found no changes to occur in litmus milk with any of these isolates. Although no mention was made regarding the reduction of nitrates by overseas workers (Baynes & Simmons, 1960; Livingston & Hardy, 1964), Worthington & Bosman (1968) found that their strain, as well as the Australian one, reduced nitrates to nitrites. However tests conducted at the Regional Veterinary Laboratory on these two strains as well as a large number of local strains showed that the reduction of nitrates was a variable characteristic and was frequently negative.

Minor and sometimes more substantial differences in the cultural characteristics obtained with bacterial isolates from cases of ovine epididymitis were no doubt responsible for reluctance on the part of some workers to identify such isolates more accurately or to associate them with A. seminis subsequently, but rather to resort to descriptive denominations such as "Gram-negative pleomorphs" (Dodd & Hartley, 1955; Ekdahl et al, 1968), "unidentified Gram-negative organisms" (Pulsford, Eastick, Clapp & Roberts, 1967) and Actinobacillus-like organisms (Ekdahl et al, 1968).

Further impetus for conducting a detailed bacteriological study of bacterial isolates obtained from ram semen was afforded by the fact that such isolates were frequently encountered where, despite their general resemblance to A. seminis, the donor rams failed to react serologically when this antigen was used. On the other hand spontaneous reactions were obtained when antigens of the specific isolates were employed in the serological test. As many of these bacterial isolates also showed cultural characteristics which closely resembled descriptions of Histophilus ovis recovered by Roberts (1956) from a case of ovine mastitis and subsequently from cases of suppurative synovitis and pyaemia in lambs in New Zealand (Kater, Marshall & Hartley, 1962) and in Australia (Hughes, Hartley, Haughey & McFarlane, 1964; Dennis, 1974; Rahaley & White, 1977) and from the semen and on autopsy the ampullae of a ram with clinical epididymitis in Australia (Claxton & Everett, 1966), the possibility that different strains of A. seminis and culturally similar but serologically non-related organisms might be involved in genital infections in South Africa, became an aspect which warranted a detailed investigation.

It was therefore decided to study the bacteriological characteristics of A. seminis, Actinobacillus-like and other Gram-negative pleomorphic bacteria associated with ovine epididymitis more closely. For this purpose cloned samples of these organisms were selected from a large number of cultures derived from the semen of a wide variety of clinically and sub clinically affected rams which showed neutrophils in their ejaculates. During the course of this study six serologically different strains of A. seminis, which had been identified in the meantime, were also included. Both standard and special bacteriological techniques were employed to establish and compare the bacteriological characteristics of the various bacterial isolates such as morphological features including spore-formation and motility, specific dye affinity, growth characteristics, nutritional and atmospheric requirements, biochemical properties, inhibition of growth and antibiotic sensitivity.

In a final series of tests some field strains of B. ovis, a renowned pathogen of the genitalia of rams, were also examined on a comparative basis.

In general all strains of organisms examined, conformed to the description of Gram-negative, non-acid-fast cocco-bacilli which tended to be arranged singly, in pairs or in short chains. They were all non-motile, non-sporulating and not encapsulated (Van Tonder, 1977b,1979c).

During the course of these investigations it became apparent that on account of their cultural and biochemical behaviour, the bacterial cultures examined could roughly be differentiated into two main groups. The one group of organisms more closely resembled A. seminis as described by Baynes & Simmons (1960), Livingston & Hardy (1964), Worthington & Bosman (1968), Van Tonder & Bolton (1968) and Watt (1970), and possibly also included the Actinobacillus-like organisms recorded by Ekdahl et al (1968) and at least some of the Gram-negative pleomorphs associated with ovine epididymitis at some stage or another (Dodd & Hartley, 1955; Ekdahl et al, 1968; Pulsford et al, 1967). The other group of bacterial isolates which showed some distinctive cultural features, more clearly identified themselves with the descriptions given to Hist. ovis (Roberts, 1956; Kater et al, 1962; Hughes et al, 1964; Dennis, 1974; Rahaley & White, 1977).

In all cultures examined, growth occurred aerobically on primary cultures but was more luxurient when cultures were incubated in a micro-aerophilic or carboxophilic atmosphere and in the latter instance still better when 20 per cent instead of 10 per cent of air was replaced by carbon dioxide.

In all instances and under all conditions of incubation, growth was always much slower for those organisms resembling Hist. ovis. This was also experienced with the vast range of 38 ordinary and enriched media which were used for primary cultivation. (Van Tonder, 1977b,1979c).

On 5 per cent horse blood tryptose or brucella agar plates bacterial colonies after 24 hours incubation, were small pin-point or slightly larger, transparent and assumed a typical "dewdrop-like" appearance. Colonies of the A. seminis type of organisms soon assumed a greyish-white appearance and became differentiated, showing umbonate greyish-white centres, transparent peripheries and entire to undulate margins. On the other hand colonies of the Histophilus-group of organisms were not only much slower in development but tended to remain transparent, dome-shaped and undifferentiated with an entire margin. Another differentiating feature was that packed cells of these organisms showed a light yellow (lemon) colour whereas in the A. seminis strains it had a greyish-white appearance. This could best be observed by holding a loopful of growth against a white back-ground. None of the isolates tested showed haemolysis on any of the blood-enriched media used.

All isolates of both groups of organisms tested were biochemically inert, producing no reactions in any of the wide range of biochemical tests performed. Occasionally some isolates of the A. seminis group of organisms showed delayed and weak acid production in some of the carbohydrate media, notably glucose, used, but these were never consistent, not even with the same isolate.

No growth or any specific reaction could be obtained on MacConkey agar or any other selective or specific media used with any of the bacterial cultures which were examined.

Nitrate reduction was mostly negative and the occasional positive reactions found were mainly caused by A. seminis-like organisms. All strains produced hydrogen sulphide but reactions with lead acetate impregnated paper strips were never pronounced.

Whereas organisms similar to A. seminis always produced catalase but no oxidase, those resembling Hist. ovis were mostly positive and occasionally negative for catalase production and usually positive and occasionally negative for oxidase production.

All strains were found to be partially resistant to certain antibiotics such as erythromycin, neomycin and kanamycin and resistant to streptomycin.

These comparative bacteriological studies revealed that the Gram-negative pleomorphic organisms isolated from the semen or affected tissues of rams suffering from clinical and sub clinical genital infection, although they all tended to be biochemically inactive, could be divided into two main groups (Van Tonder, 1979b, 1979c). The first group of organisms resembled A. seminis, as originally described (Baynes & Simmons, 1960) very closely. The other group of organisms again were very similar to Hist. ovis originally described by Roberts (1956).

From the cultural and biochemical pattern that emanated from these studies, both types of organisms can now be readily distinguished from other Gram-negative bacteria which are sometimes found in ram semen or have been incriminated as possible causes of ovine epididymitis elsewhere. These include Pasteurella (Yersinia) pseudotuberculosis (Jamieson & Soltys, 1947; Ekdahl et al, 1968), Pasteurella haemolytica and Pasteurella multocida (Cameron, 1972; Ekdahl et al, 1972; De Long, Waldhalm & Hall, 1979; Jansen, 1980), Actino-bacillus lignieresii (Laws & Elder, 1969) and Actinobacillus actinomycetem-comitans (De Long et al, 1979).

Brucella ovis (Simmons & Hall, 1953; BuddIe & Boyes, 1953), on the other hand, can easily be distinguished from all these organisms on account of its distinctive morphological, staining, cultural and some biochemical characteristics.

During the initial stages of the studies undertaken to diagnose A. seminis infection and to establish its incidence, the value of various diagnostic tests were compared. From these results it became apparent that a high percentage of rams with clinically detectable lesions from which organisms were demonstrated in semen smears or in cultures were negative when their sera were examined by the complement fixation test. Furthermore, other inconsistencies such as varying end-point titres and non-specific reactions as well as anticomplementary reactivity were often encountered in the sheep sera examined.

As various factors were considered responsible for the discrepancy between the serological test and other diagnostic procedures as well as the inconsistent reactions often obtained in sheep sera, it was decided to investigate the entire aspect, including the parameters of the complement-fixation test, in order to expose the underlying reasons for its apparent failure and to find means whereby its sensitivity and efficacy could be improved.

In the early part of this study (Van Tonder & Bolton, 1968) and in subsequent studies, a technique was used which was similar to that of Worthington & Mulders (1969) for the detection of B. ovis antibodies. It was however found that consecutive batches of extracted antigens prepared from A. seminis varied considerably in their anticomplementary activity and that the potency of these antigens was frequently too low to permit the elimination of the anticomplementary activity by simple dilution. For this reason it was considered too unreliable to use this type of antigen in an uncontrolled system in which a standard complement dilution was employed. In order to compensate for the unpredictable anticomplementary activity of extracted antigens, a combination of the techniques for the complement-fixation test described by Clapp (1955, 1961) and Clapp, Symons & Doolette (1955) for B. ovis and Baker (1967) for the Wasserman test was introduced after certain modifications (Van Tonder, 1977b). In this technique the amboceptor is titrated beforehand in the presence of excess complement, while the complement titration is undertaken in the presence of each batch of antigen as well as a non-anti-complementary negative sheep serum. When the antigen is titrated in the presence of a positive serum as well as in the final test, 1,25 MHD (Minimal haemolytic dose) of complement is always used in the serum control tubes and 1,5 MHD in the actual test tubes containing both serum and antigen.

Whereas the implementation of the modified test procedure greatly improved the consistency and repeatability of the test results when the same antigen and sera were used, variable anticomplementary activity in consecutive batches of antigen and anticomplementary and non-specific reactivity in some sheep sera still presented serious problems from time to time.

Since antigens prepared by extraction according to the method of Worthington & Bosman (1968) were found to exhibit either poor potency or alternatively they were highly anticomplementary, which in both instances rendered them unsuitable for use, means were sought whereby satisfactory antigens could be consistently prepared.

In the first of a series of investigations the effect of the preheating treatment of the harvested bacterial cells at 60° C for 30 minutes to one hour and repeated washing in suspending fluid before final heat extraction of soluble antigens was examined. These tests proved that preheating produced antigens of low titre, while extracted antigens prepared from cell suspensions, which were not preheated prior to washing, consistently yielded,. antigens of higher potency and were less anti-complementary than those antigens which were derived from the preheated cell suspensions (Van Tonder, 1977b ).

A next step was to examine the influence of various suspending fluids on the quality of antigens extracted from bacterial suspensions. A solution of physiological saline containing 0,2 per cent formalin, on account of the high anticomplementary activity of the final product, proved to be completely unsatisfactory. Physiological saline alone yielded the more potent and less anticomplementary antigen with each of the different forms of heat treatment applied when compared to saline containing 0,5 per cent phenol. The need to include an antibacterial agent in the final product necessitated the selection of phenol-saline as the preferred suspending medium, particularly when lower temperatures were employed for the extraction of antigens. Results from these comparative studies as well as specific experiments subsequently performed revealed that the lower temperatures were more suitable for the extraction of soluble antigens. The relatively insignificant differences obtained between antigens prepared by heating 60° C for 30 minutes or 100° C for 10 minutes emphasized the importance of both temperature and duration of treatment. Thus if the temperature is increased within limits the duration of exposure should be reduced (Van Tonder, 1977b).

In concluding the investigation on the production of suitable antigen from A. seminis, the additional effect of mechanical agitation of the cell suspensions on the extracted antigens obtained and the employment of bacterial cell antigens were studied. From the results obtained, there appeared to be some advantage to the incorporation of mechanical agitation in the process of preparation of heat-extracted antigens, although these antigens showed a high degree of anti-complementary activity. When both extracted and whole cell antigens were used to assay the concentration of antibodies in the sera of sheep, the results showed that there was virtually no difference between the end-point titres when either antigen was used, while known negative serum controls failed to fix complement in the presence of both these antigens. Cell antigens prepared from A. seminis strains were found to have only negligible anticomplementary activity, especially at the higher dilutions at which titrations were carried out i.e. a range of 1: 10 to 1 :70 of a 3 per cent cell suspension. This was attributed to the fact that A. seminis strains usually produced smooth colonies, the organisms were killed at 100 to 102°C for 10 minutes and that these cell suspensions were used at very high dilutions. As these cell antigens also proved to be as reliable as extracted antigens in the detection of antibodies in the sera of infected animals, cell antigens were preferred for all subsequent routine tests (Van Tonder, 1977b).

A final aspect in the conduct of the complement-fixation test which needed clarification, was the considerable incidence of anticomplementary and non-specific activity in sheep sera. Distinction between the two phenomena was based on whether inhibition of haemolysis occurred in the serum control tubes or not. Although they could have been completely unrelated, it was more likely that non-specific activity could have represented a lesser degree of anticomplementary activity, as it was more readily observed in the sera of non-infected animals where suspicious and even positive titres suddenly appeared between consecutive tests on the same specimen.

Anticomplementary activity in sheep sera already was a well-known phenomenon which could be caused by a variety of factors including bacterial contamination (Boyd, 1966; Kabat & Mayer, 1967). Non-specific reactions on the other hand, had only been recorded twice in the available literature. Gorrie (1959) incriminated bacterial contamination of the serum as the cause of these reactions in the complement-fixation test for B. avis, while Ris & Te Punga (1963) found that similar reactions encountered in an indirect haemagglutination test developed for the detection of infection by the same organism could be removed by inactivating sera at 62,5°C for 30 minutes. During the course of this investigation various factors were identified as contributory causes (Van Tonder, 1977b).

In a substantial percentage of anticomplementary sera, bacterial growth was observed despite the presence of boracic acid preservative. Many of these sera were taken under adverse field conditions or delayed in transit. Frequently unidentified Gram-negative pleomorphic psycrophylic organisms were isolated when these serum samples were inoculated onto nutrient agar plates. In order to overcome this problem, bacteria were deposited by centrifugation at 4 000 g for 20 to 30 minutes. If the sera were still uncomplementary, they were diluted with an equal volume of a 1: 10 dilution of complement, incubated at 37° C for 30 minutes and finally re-inactivated at 60° C for 30 minutes. After this treatment the majority of sera were without anticomplementary activity. Where sera were to be transported over long distances it became standard practice to inactivate each sample at 60° C for 30 minutes immediately after separating the serum from the blood clot.

Anticomplementary and non-specific activity were however also found in sera which showed no evidence of bacterial contamination. In a specific investigation, these sera were inactivated at different temperatures over the range of 56° C to 64° C for a period of 30 minutes. From these results it became apparent that temperatures of 60 to 62° C eliminated both non-specific as well as anticomplementary activity from the test sera without loss of specific fixation of complement by antibodies. Whereas the lower temperatures failed to remove the undesired qualities in the sera, temperatures in excess of 62° C applied for a period of 30 minutes not only caused a certain amount of opacity of the serum samples presumably due to some degree of coagulation of the serum proteins, but also decreased the concentration of complement-fixing antibodies to an appreciable extent. Inactivation at 60° C for 30 minutes was thereafter adopted as a standard procedure.

Despite the elimination of a large proportion of the anticomplementary and non-specific activity by heating, additional problems were encountered from time to time. These reactions suggested a low-grade anticomplementary or non-specific activity and it became evident that they were due to the combined anticomplementary effect of the antigen and the serum when added together in the actual test. These reactions were completely eliminated by the modified test procedure in which the complement was titrated and standardized in the presence of a combination of both antigen and a selective negative serum, and where 1,25 MHD of complement was added in the control tubes and 1,5 MHD in the actual test (Van Tonder, 1977b).

The employment of more potent antigens prepared from A. seminis, the pre-treatment of sera to be tested and the introduction of certain modifications to the technique of the complement-fixation test improved its reliability for detecting antibodies in the sera of affected rams. However, where preliminary tests failed to detect antibody in a certain percentage of clinical cases from which A. seminis cultures had been obtained, it appeared that serologically distinct strains of A. seminis might exist. During the routine testing of ram sera it was noted that fixation of complement could be obtained when the homologous isolate was used as the antigen. Some of these sera, however, gave no positive reaction when tested against the original reference strain of A. seminis. The strains of A. seminis recovered from such rams also failed to fix complement in the presence  of the standard positive reference serum.

The suspected existence of variant strains of A. seminis was to some extent supported by the isolation of bacterial organisms from the semen or epididymal lesions of affected rams which, on account of certain cultural differences, could not be classified as A. seminis. These included organisms described as '"Gram-negative pleomorphs" (Dodd & Hartley, 1955; Ekdahl et al, 1968), "unidentified Gram-negative organisms"' (Pulsford et al, 1967) and Actinobacillus-like organisms" (Ekdahl et al, 1968). The fact that in all these investigations bacterial cultures were often obtained from clinically affected rams that failed to react to the serological test for A. seminis, which on cultural examination closely resembled Hist. ovis (Roberts, 1956; Kater et al, 1962; Hugh et al, 1964; Dennis, 1974; Rahaley & White, 1977), also strongly indicated the possible involvement of variant strains of A. seminis or even taxonomically different types of organisms, in ovine epididymitis in this country.

Since, in the earlier studies on the bacteriological characteristics of various bacterial isolates from the semen of affected rams, some distinct cultural differences were obtained with some isolates, it was decided to establish whether antigenic differences did in fact also exist.

Throughout these studies a technique was employed whereby hyperimmune sera were prepared in rabbits against the appropriate strains of organisms to be tested. Aliquots of each specific hyperimmune serum were either left unabsorbed or were adsorbed by bacterial cells of the homologous and heterologous isolates to be compared in each series of tests (Worthington & Mulders, 1969). Antibodies were assayed in unabsorbed as well as adsorbed hyperimmune sera by the complement-fixation technique, using the antigens prepared against the various isolates included in each set of tests. The extent to which homologous antibodies to heterologous strains were also produced in the specific hyper-immune sera and the rate at which specific antibody was removed by adsorption to the bacterial cells of each isolate in turn, formed the basis on which antigenic homogeneity or disparity was formulated.

From the vast number of isolates obtained from rams with clinical epididymitis and which were also included in the bacteriological studies, a number of strains were carefully selected for inclusion into one or more of the series of experiments conducted during the course of this extensive investigation. The criteria for selecting these strains were based on cultural differences, the serological behaviour of the donor rams when tested with the standard A. seminis antigen and the degree of apparent similarity or unrelatedness observed during preliminary tests conducted with a large number of strains and their homologous convalescent antisera.

After an exhaustive series of these tests and preliminary adsorption experiments, 10 strains were selected for inclusion in complete cross-adsorption experiments. Since the isolation of A. lignieresii from the lesions of rams with epididymo-orchitis had been recorded (Laws & Elder, 1969) two strains of this organism were included, one from a young Jersey heifer and the second from an adult Merino ram, both of whom were suffering from typical clinical actino-bacillosis.

When these results became available they were carefully scrutinized and it was decided to subject six strains of A. seminis, which evidently were antigenically dissimilar, as well as the two A. lignieresii strains to a final and conclusive set of tests. Hyperimmune sera were prepared against each of these strains and portions of each serum in turn adsorbed by the homologous as well as the other strains. Unabsorbed and adsorbed samples of each serum were tested for residual antibody with antigen of each of the strains involved (Van Tonder,1977b).

From the results produced by this final experiment it became evident that the two A. lignieresii strains were antigenically different but nevertheless related, and that no relationship existed between them and the A. seminis strains tested. On the other hand these results showed a close relationship between the four strains designated T981V, 70.64, 6201 and U207 while cross-reactivity between these strains and strains 1032A and V350 was of a very low order. A systematic comparison of the antibody titre of the unabsorbed and adsorbed sera of strains T981V, 70.64, 6201 and U207 confirmed that although they possessed certain common antigens, very distinct individual antigenic differences were evident.

The reactions in the unabsorbed and adsorbed sera of strains 1032A and V350 against their respective antigens showed that there was a fairly close relationship between them, but that there were also individual antigenic characteristics.

Subsequent examinations and tests also confirmed that strains T981V, 70.64, 6201 and U207 not only resembled the original isolates of A. seminis (Baynes & Simmons, 1960; Van Tonder & Bolton, 1968; Worthington & Bosman, 1968) very closely on bacteriological and cultural characteristics, but on serological cross-examination tests also showed some close antigenic relationships with these strains.

Strains 1032A and V350, on the other hand, showed similar bacteriological and cultural properties when compared to Hist. ovis (Roberts, 1956; Kater et al, 1962; Hugh et al, 1964; Dennis 1974; Rahaley & White, 1977) and were antigenically also closely related in cross-adsorption tests carried out with some of these strains (Gumbrell & Smith, 1974; Rahaley & White, 1977).

Since genital infection in rams can be diagnosed by various methods and no single method is completely reliable for this purpose, it was decided to compare and evaluate the various tests with regard to their importance as diagnostic aids in the diagnosis of A. seminis infection, particularly as the sensitivity and reliability of the complement-fixation test during these studies had greatly been improved.

When carried out individually, each of these tests has specific limitations which can be overcome to a great extent only when they are conducted simultaneously. These tests include clinical, semen smear, bacteriological tests on semen or serological tests. As some of the tests require special laboratory facilities or transportation of animals and/or specimens over long distances, their relative importance as diagnostic aids had to be considered.

All these tests were conducted on a number of clinically and sub clinically affected as well as non-infected rams which were submitted to the Regional Veterinary Laboratory or were available at specific stud farms. The investigation was continued until such time as more than 50 animals within each category had been examined. Since the accuracy of the tests for determining A. seminis infection was evaluated, rams showing any evidence of B. ovis infection were not taken into account.

As the clinical examination merely serves to indicate the presence or absence of palpable abnormalities in the external genital organs which mostly result from genital infection, it could at best be used to identify animals falling within this category. Although with the necessary experience, the infectious type of lesions can be differentiated from non-specific lesions, and on account of its more acute nature, particularly in young rams, lesions due to A. seminis infection can fairly accurately be distinguished from the more chronic lesions of B. ovis, confirmation of the actual cause of genital lesions can only be obtained by means of other examinations and tests.

Sub clinical infection, on the other hand, can only be established on examination of semen smears and the demonstration of neutrophils and bacteria in the ejaculate. Although various staining methods can be employed, Stamp's modification of the Ziehl-Nielsen technique (Stamp, McEwen, Watt & Nisbet, 1950) was preferred since it does not only show up the presence of neutrophils but also B. ovis organisms which stain specifically with this method (Van Tonder, 1977b, 1977c). A. seminis, on the other hand, stains negatively with this technique but on account of its morphology and arrangement of the bacterial cells, and the fact that it is thus far the only other known pathogen occurring in the semen of rams in South Africa, it can be diagnosed with a fair amount of accuracy in semen smears stained by this method (Van Tonder, 1977c).

Bacteriological examination of semen specimens affords an accurate means of identifying bacterial organisms present in the semen and although this method on its own cannot be used to establish the aetiological importance of any particular isolate, the cultivation of established pathogens from semen specimens, from a diagnostic point of view, is always regarded as highly significant.

Whereas the presence or absence of lesions defines the value of the clinical examination, the usefulness of both the smear and bacteriological examinations is determined by the excretion of infected semen, i.e. the presence of neutrophils and bacteria in the ejaculate which is only possible while the epididymal duct is still patent i.e. before occlusion takes place.

The complement-fixation test which is used to show up specific antibodies in the serum of animals, is an indirect diagnostic aid employed to demonstrate systemic exposure to a specific infection and, in this case, genital infections such as A. seminis or B. ovis. It is therefore of particular value in clinical genital infection where, as a result of the local tissue reaction, there is also systemic involvement and in the case of sub clinical infections on the condition that systemic invasion of bacteria from within the epididymal tube had taken place at some stage or another. In the case of clinical epididymitis, this test has the advantage that it can be used to identify the causative organism through the presence of circulating antibodies for a considerable period of time after the bacteria, as a result of inevitable occlusion of the epididymal duct, has already disappeared from the ejaculate.

Of the rams presented at the laboratory, 83 were found to have clinical lesions. The semen smears of 68 of them (82,0%) were positive for A. seminis infection and 61 (73,4%) had positive semen cultures. In the complement-fixation test 59 (71,0%) of these rams had antibody titres to A. seminis of 1 in 20 and higher and 68 (82,0%) had titres of 1 in 10 and higher. Depending on the minimum level of antibodies considered positive, altogether 78 or 79 (94 to 95,2%) of the clinically affected rams reacted positively on one or more of these three tests (Van Tonder, 1977b).

Out of a total of 61 rams which had no lesions but were classified as sub clinically infected on semen smear examination, A. seminis was isolated from semen samples of 51 (83,6%) of them while 27 (44,0%) reacted positively to A. seminis on the serological test showing titres of 1 in 20 and higher and 36 (59,0%) showing titres of 1 in 10 and higher.

Altogether 93 of the rams examined at the laboratory had no lesions or showed any evidence of sub clinical infection on semen smear examination. Although no neutrophils or clearly visible organisms could be found on the semen smear, A. seminis was nevertheless isolated from the semen of one of these rams. Five rams (5,4%) showed antibody titres of 1 in 20 and higher and 19 rams (20,40;0) titres of 1 in 10 and higher when their serum was tested for antibodies against A. seminis.

On two stud farms with a high incidence of epididymitis in young stud rams despite the regular use of Rev 1 vaccine, 102 young ram lambs were available for examination. On clinical and semen smear examination 18 of these rams had clinical lesions, 43 were found to be sub clinically infected and the remaining 41 showed no evidence of infection by these tests (Van Tonder, 1977b).

When subjected to the other tests, 16 (88,9%) of the clinically affected young rams were positive for A. seminis infection on the semen smear examination, 15 (83,3%) had positive semen cultures and 15 (83,30;0) or 16 (88,9%) of them had antibodies to A. seminis in their serum at titres of 1 in 20 and higher or 1 in 10 and higher respectively. In all 17 (94,4%) of the clinically affected rams could be diagnosed as A. seminis infection by one or more of the other tests.

In the sub clinically affected group, i.e. positive on the semen smear examination, A. seminis was isolated from the semen of 34 (79,1%) of these young rams while 18 (41,9%) or 20 (46,5%) reacted to the serological test for A. seminis at the higher and lower ranges of antibody titres respectively.

From those rams which appeared to be free of infection on both the clinical and semen smear examinations, A. seminis was nevertheless isolated from the semen of 22 (53,6%) of them. Since 011 their semen smears there were no signs of an infective process in the genital tracts of these rams, they appeared to have harboured the infection in a latent form, which eventually proved to be associated with the mode of transmission of A. seminis infection (Van Tonder, 1977b).

From these results it became evident that the smear and bacteriological examination of semen and the complement-fixation test on the serum were equally reliable as individual tests for detecting A. seminis infection in clinically affected rams. When these results were collectively evaluated. i.e. taking into account all rams which reacted positively on one or more of these tests, it also became clear that A. seminis infection could be diagnosed in 94 or 95 per cent of clinically affected rams. The main reason for the improved supplementary effect from a diagnostic point of view, of these tests when conducted simultaneously, was to be found in the fact that specific antibodies could be detected by the complement-fixation test in the serum of rams which, on account of occlusion of the epididymal ducts, no longer excreted neutrophils and bacteria in their semen and could therefore not be diagnosed on semen examination. On the other hand again, A. seminis infection was confirmed on semen smear and bacteriological examination in those rams with acute clinical lesions where serological conversion had not yet taken place. In clinically affected rams, a combination of the semen smear examination and the complement-fixation tests is therefore preferred since these tests are most reliable, can be performed under most conditions and do not require the same extreme precautions for the collection, handling and transporation of specimens as are demanded for bacteriological examination of the semen (Van Tonder, 1977c). In rams without lesions, the presence or complete freedom of A. seminis infection can only be established with certainty by the smear and bacteriological examination of the semen. Furthermore in this group of animals and mainly in young ram lambs, a latent form of the infection was established by the bacteriological examination of semen specimens which appeared to be free of infection on semen smear examination. In both the sub clinical and latent forms of A. seminis infection the complement-fixation test proved to be completely unreliable. This was ascribed to the fact that in these instances the infection and for that matter inflammatory response in the epididymal duct, which strictly speaking can be regarded as extra-systemic, were too superficial in nature to permit systemic invasion by the bacteria and the stimulation of an immunological response.

The occurrence of a latent form of A. seminis infection in the semen and specific antibodies in the sera of some young ram lambs which had no lesions and also showed no evidence of infection in their semen smears were later proved to be associated with the method of transmission of the infection (Van Tonder, 1977b).

During all the investigations up to this stage it was shown that A. seminis infection was widely distributed, that the incidence was high and that it constituted a major problem to a large number of stud farms. It was therefore considered important to establish the natural routes of transmission as an essential requirement in the control and eradication of the disease. Experimental transmission of A. seminis to rams by the intratesticular and intra-epididymal routes had already been successfully carried out in Australia (Baynes & Simmons, 1960; Watt, Bamford & Nairn, 1970), in America (Livingston & Hardy, 1964) and in South Africa (Worthington & Bosman, 1968; Van Tonder & Bolton, 1968). In addition the intravenous and intramammary routes had also been used in the experimental infection of ewes in Australia by Baynes & Simmons (1960) and Watt et al, (1970). These studies did not, however, expose the possible routes of natural transmission and it was therefore decided to conduct a series of investigations to elucidate this aspect of A. seminis infection.

In the first of these experiments, it was merely endeavoured to confirm the infectivity of diluted ram semen infected with A. seminis when injected intratesticularly or intra-epididymally, and at the same time to affirm non-infectivity of the bacteria-free filtrate of the same specimen. Both rams inoculated with diluted infected semen intra-epididymally developed extensive lesions and were also positive to the other tests. Those injected intra-testicularly, except for a poor antibody response, remained negative but on post mortem examination the one ram did show a small abscess from which A. seminis could be isolated, at the approximate site of injection. The rams injected with bacterial-free filtrate remained negative throughout the experiment (Van Tonder, 1977b).

Since the intrapreputial route was regarded as the most common method of transmission of B. ovis infection from ram to ram via the vagina of the ewe (Hartley, Jebson & McFarlane, 1955; BuddIe, 1955; Clapp, Keogh & Doolette, 1957; Snowdon, 1958; Keogh, Doolette & Clapp, 1958) and also without mediation of the ewe (BuddIe, 1955; Clapp et al, 1957; Snowdon, 1958), it was also considered to be the most likely means of natural transmission of A. seminis infection and formed the basis of a second experiment.

In this experiment where two groups of rams were exposed by the intrapreputial instillation of two different concentrations of a suspension of A. seminis organisms, none of the animals developed clinical lesions although they all excreted A. seminis in the semen at one stage or another. The majority of these rams showed evidence of mild sub clinical infections on semen smear examination and positive complement-fixation titres. The fact that these low-grade sub clinical and latent infections gradually subsided during the observation period of 9 months without the development of lesions suggested a greater resistance against the infection in the older rams used in this experiment (Van Tonder, 1977b).

In a subsequent experiment, 7 to 8 month old Merino ram lambs were exposed to infection by various routes through the instillation or inoculation of suspensions of A. seminis organisms. The routes which were regarded as the most likely ports of entry of natural infection included the oral, rectal, conjunctival, preputial and nasal mucous membranes. As the preputial route, from a practical point of view, was still considered the most likely method of natural transmission, another group of these young rams was exposed by the repeated intrapreputial instillation of infective material at weekly intervals and on six occasions. In order to investigate the possible systemic spread of the infection to invade the genital tissues, two groups of young rams were also exposed by the intravenous and subcutaneous injection of bacterial suspensions.

None of the young rams exposed to infection by any of these routes developed the typical acute clinical lesions associated with natural A. seminis infection and neither did those repeatedly exposed via the intrapreputial mucous membrane. The majority of these rams did, however, show intermittent low-grade sub clinical or latent A. seminis infections or developed complement  fixing antibodies which seemed to confirm some degree of transmission of the infection. The failure of these young rams to develop acute, clinical lesions and the tendency to overcome the sub-clinical or latent infection with advancing age, again confirmed a high level of resistance to A. seminis infection even at the age of 7 to 8 months and the unlikelihood of these routes being involved in the natural transmission of the disease. These findings and conclusions were in fact supported by the high incidence of the natural acute clinical infection in young ram lambs within their isolated groups at or shortly after weaning under field conditions (Van Tonder, 1977b).

The rams exposed by the intravenous and subcutaneous routes showed severe systemic or local reactions while two rams in the intravenous group also developed acute purulent arthritis, which together with the isolation of A. seminis from these lesions seemed to have illustrated the invasive properties of this organism. This also appeared to be of some significance as A. seminis had been incriminated as the cause of polyarthritis in young lambs by Watt et al (1970). Furthermore H. ovis (Roberts, 1956), which was claimed by these authors to be similar to A. seminis, had also been recovered from cases of synovitis and pyaemia in lambs in New Zealand (Kater et al, 1962) and Australia (Hughes et al, 1964; Dennis, 1974; Rahaley & White, 1977).

One further experiment was conducted to evaluate repeated intranasal instillation of infective material as a possible method of natural infection. A suspension of A. seminis organisms was instilled 17 times into a nostril of each of 10 adult Merino rams on either alternate or every third day.

At no time during the period of observation were lesions of the genital organs, suggestive of infection, detected. Neutrophils were sporadically detected and on odd occasions A. seminis was recovered from the semen. In most of the rams complement-fixing antibodies were detected as soon as 13 days after exposure but, as in the previous experiments, their concentration and persistence varied from one animal to another. From the results obtained in this experiment it was concluded that adult rams were highly resistant to A. seminis infection, even when repeatedly exposed by the intranasal route which appeared to be of no importance in the natural mode of transmission of this particular infection (Van Tonder, 1977b).

During the course of these transmission experiments, another experiment was also undertaken to determine the pathogenicity of A. seminis infection to pregnant ewes exposed by various routes. Different groups of ewes which were approximately 3 to 3½ months pregnant when selected for the experiment were exposed to infection by intravenous inoculation, intranasal or conjunctival instillation and oral administration. A group of five ewes from the original flock, which were also pregnant but showed a natural infection of A. seminis when vaginal specimens were examined beforehand, were also included as part of the experiment.

With the exception of the ewes exposed by the intravenous route which showed mild transient fever reactions, none of the ewes showed any evidence of systemic infection. Vaginal swabs and colostrum samples taken at the time of lambing were negative in all cases with the exception of one ewe in the naturally infected group from which A. seminis was recovered from the colostrum.

Although none of the ram lambs born from these ewes developed clinical lesions during the course of the experiment, some of them, however, showed evidence of sub-clinical or latent A. seminis infection. As other possible means of exposure were eliminated, these ram lambs could only have acquired this form of infection from the maternal environment. The presence of A. seminis infection in the vaginal canal of the five ewes in the advanced stages of pregnancy and in the colostral milk of one of them after lambing also implicated the ewe in the natural transmission of the disease (Van Tonder, 1977b).

Following these observations, all subsequent experiments were mainly designed to investigate the role played by the ewe and to establish the importance of venereal spread in the natural transmission of A. seminis infection. In the first of these experiments two groups of ewes were exposed to the infection at the time of mating. The first group was exposed by mating to three naturally infected rams and those ewes which returned on heat during the subsequent cycle were also served by non-infected fertile rams after they had again been mated by the original infected ram, in order to ensure maximum conception and production of offspring.

Exposure in the second group of ewes was effected by the intravaginal instillation of a suspension of A. seminis organisms approximately 5 to 10 minutes prior to mating by non-infected fertile rams.

Suitable repeated tests were conducted on all the animals before the onset of the experiment to confirm the presence or absence of infection. All rams, including the vasectomized rams used, were also examined at specified intervals during the mating period and also for variable periods thereafter.

Vaginal swabs for bacteriological examination and blood for the complement-fixation test were collected from the ewes when they were approximately 2 to 3 months pregnant. From the ewes that lambed vaginal swabs and colostrum samples for bacteriological examination were collected as soon as possible after lambing. Wherever possible a complete range of specimens for bacteriological examination was collected from each lamb that was born dead or died shortly after birth. Specimens of various organs and tissues were also collected from 17 ewes in both groups which failed to lamb and were slaughtered 6 to 7 months after they were mated and exposed to infection.

The ram lambs born from these ewes were subjected to thorough clinical, bacteriological and complement-fixation tests conducted at variable intervals until the experiment was terminated when they were approximately 12 months old.

This experiment was subsequently repeated to confirm the earlier findings, with the difference that virgin Merino ewes were used. The influence of the infection on their genital tracts was also studied and the ewe lambs born from these ewes were also included in order to determine their role in the maintenance of the infection in any particular flock.

With the exception of one ram, all the non-infected fertile and vasectomized rams used in the different groups of ewes in both experiments, started to excrete A. seminis organisms in their semen or accessory fluid, at one stage or another during the mating period. These rams however did not develop lesions or excrete neutrophils in the semen and also failed to develop complement-fixing antibodies. Only one fertile ram retained the infection for longer than 10 weeks and continued to excrete A. seminis in the semen up to 6 months later. These findings again confirmed the occurrence of a latent form of the infection in rams. In the case of these rams the infection could only have been acquired from the ewes, which in turn could only have become infected through mating to naturally infected rams or by artificial exposure. Although it was again concluded that transmission of the natural disease to rams was not brought about in this way, rams acquiring a latent form of infection in this manner could nevertheless play an important role in maintaining the disease within, and also spreading it between ewe flocks during the mating period (Van Tonder, 1977b).

During these experiments it was established that A. seminis infection was not only successfully transmitted to ewes by venereal contact or artificial exposure at mating, but in addition it produced an acute inflammatory reaction in the female genital tract which lasted for approximately 3 to 4 weeks. Furthermore A. seminis could be isolated from the vaginal tract of most of the ewes up to 2 to 3 months of pregnancy and from some ewes also after lambing, which proved that the infection could be harboured in the vaginal canal throughout pregnancy. The isolation of A. seminis from various organs and tissues and also from different parts of the genital tract of four ewes that suddenly died during the course of these experiments and in various stages of pregnancy, showed that a general systemic invasion took place from the genital canal and was possibly responsible for these deaths. On the other hand systemic invasion from the genital canal and non-fatal bacteraemia as part of the natural course of the infection were also confirmed by the distinct immunological response shown by the majority of ewes after exposure during mating. The fact that A. seminis could not be isolated from colostrum samples of any of the ewes in both experiments seemed to have excluded this route as being involved in the natural transmission of the disease (Van Tonder, 1977b).

The results of the bacteriological tests conducted on lambs that died at or shortly after birth have proved beyond doubt that maternal transmission of A. seminis infection occurs very readily, as those lambs which showed the presence of the infection in various organs and also in various parts of their urogenital system could only have acquired the infection in the prenatal period. It is unlikely that the infection could have spread to this extent, and in the limited space of time, if transmission occurred in the vaginal canal during birth.

The isolation of A. seminis from various parts of the urogenital system of 8 out of 10 ewe lambs born from infected ewes in the second experiment and slaughtered when they were 5 to 7 months old, provided further incriminating evidence of the importance of venereal and maternal transmission in the occurrence of the natural disease. This, together with the fact that A. seminis was isolated from the vagina and bladder of 8 out of 17 ewes which failed to conceive and were slaughtered 5 to 7 months after exposure, also emphasized the important role played by the ewe, not only in respect of the venereal and maternal transmission of the disease, but as carriers of the infection also in the incidence and dissemination of the infection within and between flocks (Van Tonder, 1977b).

None of the ram lambs born from the infected ewes in both these experiments developed the typical acute clinical lesions. From the preputial swabs taken on four occasions at monthly intervals before weaning, A. seminis was isolated with increasing frequency. With a single exception, all the ram lambs excreted A. seminis in the semen at one stage or another. One month after weaning, the number of positive samples of A. seminis from the semen was already at a peak level in all groups and remained at this level until 6 months after weaning, when it steadily decreased until termination of these experiments. The appearance of neutrophils in the semen was a more variable feature between examinations and although up to 50 per cent of the rams showed neutrophils in their semen smears they were never in abundance. The percentage of positive semen smears showed an increase from the first to the second month after weaning after which stage, despite some fluctuations, it slowly decreased towards the end of the experimental period. On the first complement-fixation test conducted when these ram lambs were approximately 1 month old, a percentage from each group showed low antibody titres against the homologous strain of A. seminis, which however disappeared within the following month. Although these antibodies could have been the result of active stimulation and formation in the pre- and post-natal period, the possibility that at least part of it was derived from the colostrum cannot be excluded. From a month before weaning however, the incidence of complement-fixing antibodies started to increase again and rose to a peak which lasted from approximately the third to the seventh month after weaning, after which they gradually decreased.

Although none of the ram lambs born from the different groups of ewes in these experiments developed the typical acute lesions of the natural disease, the high incidence of sub clinical and latent A. seminis infection in these young rams proved beyond doubt that the infective agent was successfully transferred from the mothers to their offspring. As these rams were kept in isolation and had no contact with sheep outside their groups, they could only have acquired the infection from their mothers. Furthermore the results have proved that transn1ission of the infection could only have occurred in utero (Van Tonder, 1977b).

A striking feature observed in all the ram lambs in these experiments was the retarded development of their genitalia. This was especially noticeable in the size of the testes and the persistence of preputial adhesions during subsequent examinations. On account of the presence of these adhesions no semen could be collected from some ram lambs for a variable period after weaning, while in a large number this could only be done with great difficulty. Apart from the fact that in these cases the penis could only be extruded partly and to a variable extent, the semen was also of very poor quality, often being pure accessory fluid. These signs of retarded development were still clearly discernible in some of the ram lambs in all the different groups when they were approximately 10 months old. The less favourable nutritional conditions before and after weaning, which were reflected in the poor bodyweight and general condition of these ram lambs, were no doubt responsible for the delayed genital development. The influence of nutritional levels on the growth rate and on the development of the male reproductive organs is well-known and has been described by several workers (Johnstone, 1948; Wiggins and Terrill, 1953; Watson, Sapsford & McCane, 1956; Belonje, 1965; Pretorius & Marincowitz, 1968). It was therefore postulated that the failure of the ram lambs in these experiments to develop clinical lesions could be attributed to the delayed development of their genital organs and hence semen production, resulting in an unfavourable environment for the rapid growth and multiplication of the organisms at a stage when the incidence of sub clinical and latent infections were at their maximum. Substantiation for this belief was to be found in the results of investigations on stud farms with severe problems of A. seminis infection, where it was noticed that almost all the young rams which developed acute clinical lesions were the most well-developed and earlymaturing individuals within their age groups.

In order to verify this point of view two further investigations and one more transmission experiment were carried out. For the purpose of the first investigation, two stud farms where A. seminis infection existed and where the young rams after weaning were kept and reared in separate groups, as is customary on most stud farms, were selected. Special care was exercised to ensure that these young rams had no contact with sheep other than members within their specific age group or their mothers before weaning and were also reared in isolation within their groups after weaning. These young rams enjoyed optimal nutritional conditions at all stages during their growth and development. Altogether 102 young Dorper rams varying in age from 4 to 15 months were examined by the full range of tests. On clinical examination 18 (17,6%) were found to have lesions, of which all but two were confirmed to have been caused by A. seminis. Of the remaining rams sub clinical A, seminis infection was established in 43 (42,1 %), latent infection in another 22 (21,6%), while only 19 (18,6%) were free of any sign of infection. It was specifically noticed that of these young rams of which almost 50 per cent were between 4 and 6 months old and therefore recently weaned, those which had lesions were mostly of the younger age group and of the better grown and early maturing type (Van Tonder, 1977b).

The second investigation was conducted on a farm where severe clinical A. seminis infection was experienced annually. In sheer desperation and under the impression that his existent ram-rearing facilities constituted a concentrated source of infection, the owner erected new facilities some distance away. At weaning, the ram lambs of that particular season were classed into a group of potential stud rams which were housed in the new facilities under optimal nutritional conditions and those to be sold as flock rams which were moved to a distant camp under natural grazing conditions. Both groups of rams only had contact with other members of their age group and their mothers in the pre-weaning period and members within their respective groups after weaning.

At the time of the investigation when these rams were from 7 to 10 months old, 29 (27,9%) of the 104 of them kept at the new premises had clinical lesions, a further 39 (37,3%) were sub clinically infected and 36 (34,6%) were free of infection on semen smear examination. Of their 151 counterparts kept on natural grazing and which showed distinctly poorer body condition and genital development, only one (0,7%) had a small chronic lesion in the head of the one epididymis, 7 (4,6%) of them showed a mild sub clinical infection and 143 (94,7%) were free of any sign of infection on semen smear examination (Van Tonder, 1975).

The results of both these investigations not only supported maternal transmission of A. seminis infection very clearly, but also emphasized the importance of optimal nutritional conditions and consequent genital development in the occurrence of the natural disease (Van Tonder, 1977b).

In a natural transmission experiment, a Dorper ram with a lesion in the left epididymal tail and a severe A. seminis infection in the semen, which was still judged to be fertile, was hand-mated on two occasions to a number of ewes that proved to be free of infection when tested immediately before each mating. The ewes as well as their ram lambs after weaning were kept as separate groups and were not allowed contact with any other sheep. These animals were maintained on irrigated pastures and received a supplementary ration.

All the ram lambs showed rapid growth and early sexual development, determined by their prime condition, testicular size and early detachment of the penis. Two of these ram lambs developed acute clinical lesions when they were approximately 4 months old and another three when they were about 6 months old. Of the original 20 ram lambs that were born from the ewes in the experiment, 8 (40,0%) were free of infection and were removed when they were approximately 9 months old, while the remaining 7 (35,0%) that showed sub clinical A seminis infection were retained and retested on two occasions until the age of 11 months when the experiment was terminated. At the end of this period another one of these rams had developed a lesion in the right epididymal tail, two of them had become free of infection and four of them were still sub clinically infected. The results of this experiment again proved the importance of venereal and maternal transfer of the infection in the natural transmission of the disease and the role played by optimal nutritional conditions, culminating in early sexual development, in the development of clinical A. seminis infection (Van Tonder,1977b).

In order to study the effect of the elimination of infected breeding rams on the incidence of A. seminis, infection, the results of the annual tests on all sale rams of a particular stud farm before and after the introduction of a strict programme of regular testing and elimination of A. seminis-infected breeding rams were compared.

From the second year after the introduction of this programme i.e. when the first male progeny of the tested breeding rams became available for testing, the incidence of clinical epididymitis in the Merino sale rams which varied from 4 to 12 per cent in the preceding five years was reduced to 1 to 2 per cent in the subsequent 6 years, while the incidence of sub clinical infection which varied from 12 to 35 per cent in the preceding period dropped to an average of 4 per cent in the subsequent period. In the Dorper sale rams, the clinical incidence of A. seminis infection which ranged from 8 and 10 per cent before introduction of the breeding programme was reduced to nil and the sub clinical incidence from 16 to 28 per cent to an average of 4 per cent in the subsequent 6 years. The failure to completely eliminate the infection on this particular farm within the period of investigation was attributed to the presence of infected and carrier ewes and the mechanical transfer of the infection by the rams through venereal contact during mating. It nevertheless became evident that although carrier ewes and rams can playa role in the distribution and incidence of A. seminis infection, the ideal circumstances for the natural transmission of the disease would include the presence of rams with infected semen (Van Tonder, 1977b ).

From the information obtained from these experiments and investigations it was finally concluded that the mechanism involved in the natural transmission and course of A. seminis infection follows a pattern of venereal transmission from ram to ewe by infected semen and intra-uterine transmission from the ewe to the lamb after a systemic spread from the vagina during the course of pregnancy. The ram lambs will be born with or without the, infection which can be harboured under optimal conditions in the genital tract in a sub clinical or latent form. Depending upon the nutritional status and resulting early sexual development, susceptibility and the degree of sub clinical infection, some ram lambs will develop acute clinical lesions at puberty or at some early stage of their life but usually before the two-tooth stage. The majority of ram lambs, however, tend to overcome the sub clinical or latent infection with advancing age, while a small percentage remains sub clinically infected for the rest of their lives. Occasionally some of these rams may develop chronic epididymal lesions (Van Tonder,1977b).

Ewe lambs, on the other hand, may also be born harbouring the infection in their urogenital system for a period of 5 to 7 months. In a subsequent experiment such ewe lambs were shown to maintain the infection in the vaginal canal for up to 22 months of age, despite the fact that they had no contact with any other sheep at any stage or with their mothers or male counterparts in the post-weaning period (Van Tonder, 1978). Ewe lambs infected in this way may therefore act as carriers and play an important role in maintaining and disseminating the infection within a flock.

Concluding the epizootological studies, an evaluation was made of the possible influence of A. seminis infection on fertility and reproduction. In order to do this, accurate records of the semen quality of the different rams used and the breeding performance of the ewes included in the maternal transmission experiments were kept.

When the conception figures on first service, i.e. ewes which did not return on heat during the second oestrous cycle, were compared between those ewes served by different naturally infected rams and the ewes that were mated by non-infected fertile rams after artificial exposure, it became evident that A. seminis infection exerted a variable adverse effect on semen quality and hence the fertilizing capacity of a ram. The poor conception figure (19%) obtained with one of the naturally infected rams with heavily infected and poor quality semen proved that the infection could have a severe influence on fertility with consequent low conception and lambing percentages. On the other hand, the comparative figures produced by some of the other naturally infected rams did not show a marked influence of infection on conception in spite of the presence of large numbers of neutrophils in their semen. From these results it became evident that A. seminis infection in the semen only had a noticeable adverse influence on fertility and conception once the severity of the infection and deterioration of the semen quality have reached a certain critical stage. Once this stage is reached, the adverse effect will increase proportionally to the increase in severity of the infection and deterioration of the semen (Van Tonder, 1977b).

During these examinations, it was also found that the relationship between the motility rating judged by the hanging-drop method and the percentage of viable sperm cells as determined by the eosin-nigrosin smear examination did not conform to the normal standards (Starke, 1949; Van Rensburg, 1973). This was attributed to the influence of infection on the viability of the spermatozoa and an increase in their fragility when subjected to the more severe treatment of the eosin-nigrosin staining method. A large number of spermatozoa which were still viable and actively motile under the more favourable conditions prevailing during the examination of a hanging drop appeared to have developed secondary abnormalities such as curled tails or detached heads and consequently died off during the preparation of eosin-nigrosin smears. In practice this would amount to a ram with infected semen performing somewhat better during the natural mating process where the semen is deposited in the favourable vaginal environment that could be predicted by the hudgement of a semen specimen by either of these methods (Van Tonder, 1977b).

In the transmission experiments a certain percentage of ewes in both groups, i.e. those mated to infected rams or to non-infected fertile rams after artificial exposure to A. seminis infection, did not show recurrent oestrus but nevertheless failed to lamb. Although this could have been due to other causes, it appeared more likely that a uterine infection resulting in a persistent corpus luteum (Condert & Short, 1966) or early embryonic death was responsible. The tendency for this percentage of ewes to have been higher in those mated by infected rams was most probably due to the infection reaching the uterine lumen more readily when present in infected semen than when deposited in the vagina in the form of an artificial culture of non-motile bacteria (Van Tonder, 1977b).

Since a variety of causes could have been involved which could not be eliminated altogether, the role played by A. seminis as a cause of early embryonic death, abortions, stillbirths or neonatal mortality could not be assessed with accuracy in these experiments. The high percentage of stillbirths and neonatal deaths encountered in both the naturally or artificially exposed ewes and the fact that A. seminis was recovered from specimens of some of these lambs, however, seemed to have indicated that this infection was responsible for at least some of these deaths (Van Tonder, 1977b). This view was at least supported by the findings of Baynes & Simmons (1966), who produced stillbirths in a ewe experimentally infected during the advanced stages of pregnancy.


Sustained research efforts

In addition to the regular examination of breeding rams and the elimination of rams that showed any sign of infection, other possible means of controlling A. seminis infection were investigated during the course of the epizootological studies and are still being examined.

At first all efforts were aimed at the development and formulation of effective methods of treatment, particularly of sub clinically infected cases. In the initial experiments various groups of carefully selected sub clinically infected rams were subjected to different treatment regimes employing a wide variety of antibiotics and dosage levels, administered by intramuscular injection. All the animals included in all these trials were subjected to a full range of tests on at least two to three occasions at least 3 days apart before the onset of the various treatments. The degree of infection in each ram was also monitored at frequent regular intervals during the period of observation which varied from 6 to 24 weeks. None of these treatments proved beneficial in any way. After further repeated attempts at treatment, using a wide variety of proprietary antibiotic products this form of systemic treatment was abandoned. Failure of systemic treatment was particularly evident when the intensity of sub clinical infection remained unchanged in a group of rams which were given 7 consecutive courses of different antibiotic treatments, consisting of 6 daily systemic injections of each. Since all the antibiotics used proved highly effective in in vitro tests, the only conclusion that could be arrived at was that, even at the high dosage levels employed in some of these trials, the transfer of these substances from the circulation to the testes and epididymides was insufficient to have a bactericidal and curative effect (Van Tonder & Bolton, 1975). On account of these observations as well as the results obtained in some vaccination trials, the possible existence of a genital barrier was postulated (Van Tonder & Bolton, 1976). This barrier was contemplated to act as a protective mechanism and to operate in a selective manner in order to prevent possible harmful substances from entering the testicular and epididymal tissues and to exert their detrimental influence on an extremely sensitive product such as semen.

In the meantime it was also endeavoured to develop experimental vaccines to be used on a random basis. This was mainly because A. seminis infection could not be reproduced by exposure via the normal contact routes, which made laboratory evaluation procedures of vaccines impossible. All experimental vaccines consisted of water-in-oil emulsions of killed bacteria and, depending on what strains were involved on the farms included in these trials, either polyvalent or monovalent vaccines were used. In the first series of trials half the number of ram lambs born from studs on a number of infected properties were vaccinated at various ages in the pre-weaning period, receiving two injections 3 weeks to 1-month apart, while the other half were left as untreated controls. In the subsequent series of experiments half the number of stud ewes on some farms were also vaccinated twice during the advanced stages of pregnancy. The ram lambs born from both the vaccinated and non-vaccinated ewes on some of these properties were also either vaccinated or left as untreated controls.

In all instances all the animals included in these trials were serologically tested before, and at suitable intervals after vaccination in order to monitor the generation of complement-fixing antibodies. The ram lambs used in these experiments were subjected to whatever tests were possible at any given stage and to the full range of clinical, semen smear, bacteriological and serological tests whenever this could be accomplished.

Altogether five different series of inoculation trials were conducted over a period of 4 years. None of these trials however showed any beneficial effect by the vaccines and methods followed, since no differences in the incidence of clinical, sub clinical and latent A. seminis infection could be established between vaccinated and non-vaccinated ram lambs or those born from vaccinated or non-vaccinated ewes (Van Tonder & Bolton, 1978).

As all these experiments high specific antibody titres were elicited in the vaccinated ram lambs, it became evident that these humoral antibodies had no effect on the development of the natural infection which again appeared to have been due to the fact that the infection had already reached its destination when they became available. The fact that these humoral antibodies also had no effect on the sub clinical and latent infection in the genital tract, amongst other possible explanations, also suggested their inability to reach the site of infection, possibly as a result of an effective blocking mechanism.

In view of the earlier supposition of the existence of some form of a genital barrier which became evident during the treatment trials, as well as the presumption that circulating antibodies were effectively prevented from entering the tubular genital system, one pilot experiment was also conducted to verify the last-mentioned possibility. Specific immuno-globulin (Ig G) was prepared against one specific strain of A. seminis, labelled with radio-active iodine (I125) and injected intravenously into each of three rams, while a fourth ram which acted as a control animal, was injected with unbound iodine only. The concentration of radio-active particles in the blood, urine and semen of these rams was monitored before treatment and at frequent intervals during the day for the first 4 days, once daily for 14 days and once every 3 days on another four occasions after treatment.

From these results it could be seen that the blood concentration of radio-active iodine rapidly declined during the first three days and continued to do so at a very reduced rate until the end of the period of observation when it virtually disappeared. The high concentration of radio-active particles in the urine of the three rams injected with labelled immunoglobulin during the first three days after treatment and the steady decline towards the end of the period seemed to have indicated that most of the labelled globulin was excreted in this way. The much lower urine values obtained in the control animal were probably due to the fact that a large proportion of the unbound iodine was cleared from the bloodstream in some other way. In all these animals the semen concentration of radio-active particles was of no significance at all times and when (present was regarded as merely due to urethral contamination by urine. It was therefore concluded that the transfer of immuno-globulin (lg G) when bound to radio-active iodine (I125) and unbound iodine (1125) by itself from the blood circulation to the genital tissues and hence their excretion in the semen was successfully prevented by some or other mechanism. Although the greater molecular size of the substances used in this experiment could have played an important role, some evidence to the effect that an effective genital barrier might indeed exist had nevertheless been produced (Van Tonder & Bolton, 1976). Support for the belief in the possible presence of such a protective mechanism in rams was subsequently found in a report defining a physiological blood testis barrier in rats and monkeys (Osler, 1978).

In spite of the negative outcome of previous treatment trials, a series of experiments were again carried out more recently. Several groups of ram lambs varying in age from 5 to 8 months and suffering from sub clinical and latent A. seminis infection were subdivided into subgroups and subjected to various antibiotic treatments, while suitable untreated controls were included each time. In each experiment the ram lambs were examined beforehand and the degree of infection rated according to the concentration of neutrophils and bacteria in their semen. The ram lambs were then allocated to the different treatment and control groups in such a way that the different degrees of infection were equally represented within each group.

At the termination of these experiments when the ram lambs were approximately 12 to 13 months old, there were no significant differences in the percentage of the young rams in the treatment groups, which became free of the infection when compared to the control groups. The percentage of ram lambs which got rid of the infection within the period of treatment and observation varied from 70 to 1000;0 in both the treated and untreated animals and the difference in this percentage between the treated and untreated rams in each experiment never exceeded 3 per cent either way. These results not only supported the previous findings with regard to the ineffectiveness of systemic antibiotic treatment but also substantiated the conclusions drawn from earlier work that sub clinically infected young rams tend to rid themselves of the infection with advancing age (Vermeulen, Kellerman & Van Tonder, 1986).


Present and future approaches

From the information presented it will be quite clear that many of the epizootological features of A. seminis infection in sheep had been unravelled and that at least the framework of mechanisms involved in the development of the natural disease had been defined.

However, there are, no doubt, many concealed facts left to be unveiled in the presentation and occurrence of this complicated and multifaceted natural infectious disease.

Present and future research will therefore be directed to this end and answers will have to be found on questions such as the exact time of transfer of the infection from the vaginal canal to the unborn lamb, the stage of transfer of the infection to the genital tract of the ram lamb, all factors involved in the development of clinical and sub clinical genital infection in the ram, the mechanisms involved in the overcoming of sub clinical infection, the precise role of carrier ewes and ewe lambs and the absolute definition of the genital barrier and its exploitation in the control of the disease.

All efforts will therefore be aimed at finding suitable control measures, most probably in the form of preventative vaccination of either one or both sexes, in which case the stimulation of some form of cellular immunity will be a likely approach.



BAKER, F.J., 1967. Handbook of Bacteriological Technique, 2nd ed. London: Butterworths.

BAYNES, I.D. & SIMMONS, G.C., 1960. Ovine epididymites caused by actinobacillus seminis N.Sp. Australian Veterinary Journal 36 : 454 - 459.

BAYNES, I.D. & SIMMONS, G.C., 1966. Actinobacillus seminis in sheep. Australian Veterinary Journal, 45,225.

BELONJE, P .C., 1965. Observations on the post-natal development of the penis in Merino ram lambs and wethers: The possible relationship to the passage of urinary calculi. Journal of the South African Veterinary Medical Association, 36, 381 - 383.

BOYD, W.C., 1966. Fundamentals of Immunology 4th ed. New York: John Wiley & Sons.

BUDDLE, M.B., 1955. Observations on the transmission of Brucella infection in sheep. New Zealand Veterinary Journal, 3, 10 - 19.

BUDDLE, M.B. & BOYES, BETTY, W., 1953. A Brucella mutant causing genital disease of sheep in New Zealand. Australian Veterinary Journal, 29, 145 - 153. CAMERON, C.M., 1972. Immunization of sheep against Pasteurellosis. Journal of the South African Veterinary Association, 43,77 - 81.

CLAPP, K.H., 1955. A complement fixation test for the diagnosis of ovine brucellosis with special reference to epididymitis. Australian Veterinary Journal, 31, 27 - 28.

CLAPP, K.H., 1961. A comparison of various antigens used in the complement fixation test for ovine brucellosis. Australian Veterinary Journal, 37,188 - 190.

CLAPP, K.H., KEOGH, J. & DOOLETTE, J.B., 1957. Ovine brucellosis. 19th Annual Report, Institute of Medical & Veterinary Science, South Australia, 61 - 65.

CLAPP, K.H., SYMONS, L.E.A. & DOOLETTE, J.B., 1955. The application of a complement fixation test to the diagnosis of ovine brucellosis. Australian Veterinary Journal, 31,29-30.

CLAXTON, P.D. & EVERETT, R.E., 1966. Recovery of an organism resembling Histophilus ovis from a ram. Australian Veterinary Journal, 42, 457 - 458.

COUDERT, S.P. & SHORT, R.V., 1966. Prolongation of the functional life of the corpus luteum in sheep with experimental uterine infections. Journal of Reproduction and Fertility, 12, 579 - 582.

DE LONG, W.J., WALDHALM, D.G. & HALL, R.F., 1979. Bacterial isolates associated with epididymitis in rams from Idaho and Eastern Oregon flocks. American Journal of Veterinary Research. 40, 101 - 102.

DENNIS, S.M., 1974. Perinatal lamb mortality in Western Australia 3. Congenital infections. Australian Veterinary Journal, 11, 507 - 510.

DODD, D.C. & HARTLEY, W.J., 1955. A specific suppurative epididymitis of rams. New Zealand Veterinary Journal, 3,105-110.

EKDAHL, M.O., MONEY, D.F.L. & MARTIN, C.A., 1968. Some aspects of epididymitis of rams in New Zealand. New Zealand Veterinary Journal, 16, 81 - 82.

ELBERG, S.S. & FAUNCE, K., 1957. Immunization against Brucella infection, VI. Immunity conferred on goats by a non-dependent mutant from a streptomysin-dependent mutant strain of Brucella melitensis. Am. J. Path. 31: 1065-1075.

GORR1E, C.J.R., 1959. Diagnosis of ovine brucellosis. Australian Veterinary Journal, 35, 500 - 501.

GUMBRELL, R.C. & SMITH, J.M.B., 1974. Desoxyribonucleic acid base composition of ovine actinobacilli, Journal of General Microbiology 84, 399 - 402.

HARTLEY, W.J., JEBSON, J.L. & McFARLANE, D., 1955. Some observations on the natural transmission of ovine brucellosis. New Zealand Veterinary Journal, 3, 5 - 10.

HERZBERG, M., ELBERG, S.S. & MEYER, K.F., 1953. Immunization against Brucella infection. II. Effectiveness of a streptomycin-dependent strain of Brucella melitensis. Journal of Bacteriology 66, 600 - 605.

HUGHES, K.L., HARTLEY, W.J., HAUGHEY, K.C. & McFARLANE, D., 1964. A study of perinatal mortality of lambs from the Oberon, Orange and Monaro districts of New South Wales. Proceedings of the Australian Society for Animal Production,5, 192.

JAMIESON, S. & SOLTYS, M.A., 1947. Infectious epididymo-orchitis of rams associated with Pasteurella pseudo tuberculosis. Veterinary Record, 59, 351 - 353.

JANSEN, B.C., 1980. The aetiology of ram epididymitis. Onderstepoort Journal of Veterinary Research, 47, 101-107.

JOHNSTONE, I.L., 1948. The growth and development of the penis in sheep: Their possible relationship to posthitis. Australian Veterinary Journal, 24, 86 - 88.

KABAT, E.A. & MAYER, M.M., 1967. Experimental lmmuno-chemistry, 2nd ed. Springfield: Charles C Thomas.

KATER, JOAN, C., MARSHALL, S.C. & H~RTLEY, W.J., 1962. A specific suppurative synovitis and pyaemia in lambs. New Zealand Veterinary Journal 10, 143 - 144.

KEOGH, J., DOOLETTE, J.B. & CLAPP, K.H., 1958. The epidemiology of ovine brucellosis in South Australia. Australian Veterinary Journal, 34, 412 - 417.

LAWS, L. & ELDER, JEAN, K., 1969. Mastitis in sheep caused by Actinobacillus lignieresii. Australian Veterinary Journal, 45, 401 - 403.

LIVINGSTON, C.W. & HARDY, W.T., 1964. Isolation of  Actinobacillus seminis from ovine epididymitis. American Journal of Veterinary Research, 25, 660 - 663.

OSLER, G., 1978. Research review: The blood testis barrier in rats and monkeys. Research in Reproduction, 10, 4.

PRETORIUS, P.S. & MARINCOWITZ, G., 1968. Post-natal penis development, testes descent and puberty in Merino Iambs on different planes of nutrition. South African Journal of Agricultural Science, 11,319-334.

PULSFORD, M.F., EASTICK, B.C., CLAPP, K.H. & ROBERTS, R., 196 7. Traumatic epididymitis of Dorset Horn and Poll Dorset rams. Australian Veterinary Journal, 43, 99 - 101.

RAHALEY, R.S. & WHITE, W.E., 1977. Histophilus ovis infection in sheep in Western Victoria. Australian Veterinary Journal. 53, 124 - 127.

RIS, D.R. & TE PUNGA, W.A., 1963. An indirect haemagglutination test for the detection of Brucella ovis antibodies: 1. Development of the test. New Zealand Veterinary Journal,11,94-97.

ROBERTS, D.S., 1956. A new pathogen from a ewe with mastitis, Australian Veterinary Journal, 32, 330 - 332. SIMMONS, G.C. & HALL, W.T.K., 1953. Preliminary studies on the occurrence and pathogenicity of a Brucella-like organism. Australian Veterinary Journal, 29, 33 - 40.

SNOWDON, W.A., 1958. The epidemiology of ovine brucellosis in South Australia. Australian Veterinary Journal, 34, 1 417-423.

STAMP, J.T., McEWEN, A.D., WATT, J.A.A. & NISBET, D.I., 1950. Enzootic abortion in ewes. 1. Transmission of the disease. Veterinary Record, 62, 251 - 254.

STARKE, N.C., 1949. The sperm picture of rams of different breeds as an indication of their fertility. 11. The rate, of sperm travel in the genital tract of the ewe. Onderstepoort Journal of Veterinary Research, 22, 415 - 525.

V AN DRIMMELEN, G.C., 1960. Control of brucellosis in sheep and goats by means of vaccination. Journal of the South African Veterinary Medical Association, 31, 129 - 138.

VAN HEERDEN, K.M., 1964. Results obtained by the use of Rev 1 vaccine in sheep against infectious infertility suspected to be caused by ovine brucellosis. Bulletin Office. Internationale des Epizootique, 62, 997 - 1002.

VAN HEERDEN, K.M. & VAN RENSBURG, S.W.J., 1962. The immunization of rams against ovine brucellosis. Journal of the South African Veterinary Medical Association, .33, 143 - 148.

VAN RENSBURG, S.W.J., 1973. Breeding Problems and Artificial Insemination. 2nd ed. Pretoria: J.L. van Schaik (Ltd).

VAN RENSBURG, S.W.J., VAN HEERDEN, K.M., LE ROUX, D,J. & SNUDERS, A.D., 1958. Infectious infertility in sheep. Journal of the .South African Veterinary Medical Association, 29, 223 - 233.

V AN TONDER, E.M., 1973. Infection of rams with Actinobacillus seminis. Journal of the South African Veterinary Association, 44, 235 - 240.

VAN TONDER, E.M., 1975. Unpublished data. Regional Veterinary Laboratory, Middelburg, Cape.

VAN TONDER, E.M., 1977. Unpublished data. Regional Veterinary Laboratory, Middelburg, Cape.

VAN TONDER, E.M., 1977. Actinobacillus seminis infection in sheep in South Africa. DV Sc. thesis. University of Pretoria, Pretoria.

VAN TONDER, E.M., 1977. Examination of rams for genital soundness. Journal of the South African Veterinary Association,48, 267 - 272.

VAN TONDER, E.M., 1978. Unpublished data. Regional Veterinary Laboratory, Middelburg, Cape.

VAN TONDER, E.M., 1979a. Actinobacillus seminis infection in sheep in the Republic of South Africa. I. Identification of the problem. Onderstepoort Journal of Veterinary Research, 46, 129 - 133.

VAN TONDER, E.M., I979b. Actinobacillus seminis infection in sheep in the Republic of South Africa. II. Incidence and geographical distribution. Onderstepoort Journal of Veterinary Research. 46,135 - 140.

VAN TONDER, E.M., 1979c. Actinobacillus seminis infection in sheep in the Republic of South Africa. III. Growth and cultural characteristics of A. seminis. Onderstepoort Journal of Veterinary Research, 46,141 - 148.

VAN TONDER, E.M. & BOLTON, T.F.W., 1967. Unpublished data. Regional Veterinary Laboratory, Middelburg, Cape.

VAN TONDER, E.M. & BOLTON, T.F.W., 1968. Epididymitis in rams caused by Actinobacillus seminis. Journal of the South African Veterinary Medical Association, 39, 87 -90.

VAN TONDER, E.M. & BOLTON, T.F.W., 1975. Unpublished data. Regional Veterinary Laboratory, Middelburg, Cape.

VAN TONDER, E.M. & BOLTON, T.F.W., 1976. Unpublished data. Regional Veterinary Laboratory, Middelburg, Cape.

VAN TONDER, E.M., BOLTON, T.F.W. & KELLERMAN, G.E., 1978. Unpublished data. Regional Veterinary Laboratory, Middelburg, Cape.

VAN TONDER, E.M., WORTHINGTON, R.W. & MüLDERS, MARIA, S.A., 1972. Effect of Rev. I. vaccination. Journal of the South African Veterinary Medical Association, 43,417-418.

VERMEULEN, S.O., KELLERMAN, G.E. & V AN TONDER, E.M., 1986. Unpublished data. Regional Veterinary Laboratory, Middelburg, Cape.

WATSON, R.H., SAPSFORD, C.S. & McCANE, I., 1956. The development of the testis, epididymis and penis in the young Merino ram. Australian Journal of Agricultural Research, 7, 574 - 590.

WATT, D.A., 1970. Investigations of ovine brucellosis in Merino rams of Western Australia. Australian Veterinary Journal,46, 506 - 508.

WATT, D.A., BAMFORD, V. & NAIRN, M.E., 1970. Actinobacillus seminis as a cause of polyarthritis and posthitis in sheep. Australian Veterinary' Journal, 46, 515.

WIGGINS, E.L. & TERRILL, C.E., 1953. Variation in penis development in ram Iambs. Journal of Animal Science, 12,524 - 535.

WORTHINGTON, R.W. & BOSMAN, P.P., 1968. Isolation of Actinobacillus seminis in South Africa. Journal of the South African Veterinary Medical Association, 39, 81-85.

WORTHINGTON, R.W. & MüLDERS, MARIA, S.A., 1969, Antigenic Relationship of Brucella ovis to Brucella abortus and Brucella melitensis using the complement fixation test. Onderstepoort Journal of Veterinary Research, 36, 191 - 198.

WORTHINGTON, R.W., VAN TONDER, E.M. & MüLDERS. MARIA, S.A., 1971. The incidence of Brucella ovis infection in South African rams: A serological survey. Journal of the South African Veterinary Medical Association, 43, 83 -85.



Karoo Agric 3 (9), 21-37