^{page 342 note 1} We occasionally meet with verb-forms in final i-' that might at first sight seem to contradict the statement that the termination-' is never found after the short final i of verbs tertiae infirmae. In Ebeling, op. cit., 317,Google Scholar 10, we have, for example, a 2nd masc. sing. preterite written tad-ki-', whereas in ibid., 25, 6, it is written ta-ad-ki. Since these two forms are masculine the i must be etymologically short; so the -' in tad-ki-' might seem to violate our principle. This is only apparent. For whereas ta-ad-ki is according to its context indicative, tad-ki-' is in a context that obviously requires the subjunctive (ša-pi-ri-u-tu ša tad-ki-'). Therefore tad-ki-' must be understood as a secondary formation to the older contracted tad-ku-u (ibid., 48, 25), which is due to analogy with the indicative; that is to say, tad-ki-' ═ tadki + '(u), the subjunctive afformative u being sufficiently indicated by-'; just as it is in forms like il-lak-' ═ illaku (see above). Similarly, in ibid., 58, 28, we have a 3rd masc. plur. permansive written mi-si-', in contradistinction to the singular mi-si (ibid., 17, 17), which is a secondary formation to the older contracted mi-su-u (passim); that is to say, mi-si-' ═ misi + '(u), like the common e-tir-' ═ etiru. Such secondary uncontracted forms are also found spelt somewhat differently, e.g. for iqabbū we find i-qab-bi-'-u and i-qab-bi-a; for iqabbā,i-qab-bi-'-a and i-qab-bi-a (cf. Ylvisaker, § 30). The existence of such secondary uncontracted forms need not surprise us; for exactly parallel phenomena are found in other Semitic dialects (see Brockelmann, op. cit., i, § 271). A full discussion of such forms will be published elsewhere.