As for karaiḥ in 649c, as it is intended as 'hands', apparently of the personal Krishna, a strict application of Pāṇini (AA 1.4.22 dvyekayor dvivacanaikavacane) would inevitably lead to the form karābhyām, which, however, would give a hypermetric pāda of 9 syllables, and hence a conflict with the principle api māṣaṁ maṣaṁ kuryāt chandobhaṅgaṁ na kārayet, which, although it is not Pāṇini, is rather strictly followed in classical poetry. Since taptam corresponds to the fire in pāda a, one could have a pāda d corresponding more explicitly with the moon in pāda b: Krishna brings peace through the beams of the moon, the Soma-beams: aṁśubhiḥ, which, as 'ṁśubhiḥ, would fit the metre perfectly. An objection could then nevertheless be that this gives an undesirable mixture of styles: everywhere expressions that are beautiful because of their simplicity, and here suddenly a 'learned' expression aṁśubhiḥ as moon- or Soma-beams. At the end, we can, then, leave the verse as created by the Poet as it is, and only change the translation: instead of "with his own hands" the translation of karaiḥ should be "with is own 'hands', i.e. moon-rays": the plural itself indicates that kara stands not for 'hands' but for the rays, in this case of the moon in pāda b. 

Best, Jan

aṁśubhiḥ 

 

--