Discussion 



Several potential sources of error in the test were revealed by the check tests that were not covered adequately 

 in the protocol (l). 1, 2 There is an optimum amount of moisture that the plates should contain when they are used. 

 If there is an excess of moisture, particularly if it is visible in the wells, the plate must be dried before use. Therefore, 

 the following change is recommended in the protocol. After the 1-percent layer is poured the plates should be 

 allowed to dry 1 to 2 hours with the lids off at room humidity before the reagents are added to the wells. It is still 

 recommended that a moist chamber be used for incubation, particularly in the dryer climates. It has been noted that 

 if the plates become too dry the lines are weaker so the plates should not be allowed to set for long periods before 

 being used. Each laboratory will probably have to adjust its method a little to compensate for the humidity in the 

 laboratory room. The amount of moisture can be gauged by the speed with which the reagents evaporate. Also if the 

 sides of the wells do not remain perpendicular after being cut there is an excess of moisture in the agar. 



Care must be taken to insure that the wells are completely filled. To insure getting adequate serum in the well, 

 it should be filled level with the agar surface without leaving any meniscus. 



Some laboratories did not detect some of the strong positive samples. If the control line stops at some distance 

 from the unknown well the sample must either be a strong positive or it is an unsatisfactory test and should be 

 repeated. If the control line is clearly visible and goes into an adjacent negative serum well the unknown sample must 

 be a strong positive even though there is no line of identity (fig. 1). This finding can be confirmed by making 

 doubling dilutions of the unknown serum. A line of identity should appear with the diluted serum. 



§§g 



Figure 1 



The weak positive serums were missed most frequently. Any plate not containing a good control line should be 

 discarded and the test repeated. With many of the check serums that were very weak the control precipitin lines only 

 hooked at the edge of the test serum well. If the control line is weak and does not extend to the well the weak 

 positive sample cannot be detected. It has been our experience that these serums are the ones that cause conflicting 

 results between laboratories. This does affect confidence in the test, even though these horses may be of doubtful 

 epizootiological importance. 



In a few cases, a sample that produced a strong line of non-identity was called positive. This is another case 

 where a strong control line is required. The angle formed between a line of non-identity and the positive control line 

 is different than that formed by a line of identity (fig. 2). 



Figure 2 



Numbers in parentheses refer to References at the end of this report. 

 2 "Protocol for the Immunodiffusion (Coggins) Test for Equine Infectious Anemia."' Prepared by James E. Pearson, DVM, in 

 collaboration with the Committee on Equine Infectious Anemia of the American Association of Veterinary Laboratory Diagnosti- 

 cians, E. A. Carbrey, Chairman, D. Barnett, L. Coggins, J. R. Gorham, W. L. Kadel, W. W. KirkhamJ. E. Pearson, E. Roth (October, 

 1971). 



18 



