348 BACTERIOLOGY OF THE EYE 



proof that the case is not tubercular. I have repeatedly inoculated 

 the anterior chamber, without result, from conjunctival and lacrymal 

 sac tuberculosis, which, clinically and microscopically, were typical. 



In such cases either the bacilli are too scanty in number, or their 

 virulence is too slight, or their pathogenicity for the animal is in- 

 sufficient to produce infection. 



2. EXAMINATION FOE THE TUBERCLE BACILLI. 

 (Technique, cf. p. 16.) 



We hardly get the opportunity of examining the conjunctival secre- 

 tion for tubercle bacilli in the same way as the sputum. The secretions 

 from the various forms of conjunctival tubercle, which in this respect 

 resemble lupus, are shown by the microscope not to be rich in demon- 

 strable bacilli; tears still further dilute them. The examination of 

 the secretion for the bacilli cannot be recommended, and is never made. 

 A diagnostic excision gives far better results in doubtful cases of 

 tubercle of the lids and conjunctiva. 



As Morax has shown, the pus from the lacrymal sac can show 

 tubercle bacilli, and thus a tubercular dacryocystitis can be diagnosed. 

 Such an examination can be thoroughly recommended in doubtful 

 cases. We cannot expect to demonstrate the bacilli in every case. 

 Granulation forms occur in which there is hardly any secretion. 1 

 When a secretion is present, and no bacilli are found, it would be very 

 risky to exclude tuberculosis. 



Pus from a tubercular caries (malar bone), like other cold abscesses, 

 may not give any definite result. 



In ophthalmology stained smear preparations are of little value. 



The demonstration of the bacilli in sections (cf. ' Technique,' p. 16) 

 often presents great difficulty, as most of the cases of tuberculosis of 

 the eye, especially when they are available for examination, are very 

 poor in bacilli. Examination of serial sections will usually give a 

 positive result ; the examination of a small number of sections is not 

 sufficient. 



When acid-fast bacilli are found in the superficial layers of a super- 

 ficial process, the possibility that they are ' tuberculoid ' must be 

 considered that is, that they are allied bacilli not etiologically 

 identical (cf. ' Technique,' p. 17). But when the bacilli are found in 

 tissues which histologically show typical tubercular lesions, they may 

 be considered to be tubercle bacilli. 



1 Cf. SioL Klin., 1907, p. 344, ' Haut-tuberculose. ' 



