to this rule also, for bases with short अ a, beginning and ending with a consonant, may optionally take Vṛiddhi (Pâṇ. VII. 2, 7). This option is afterwards restricted, and roots with short अ a, beginning with a consonant and ending in र् r, like a जागर् jâgar, have no option left, but are restricted afresh to Vṛiddhi (Pâṇ. vii. 2, 2). However, even this is not yet the final result. Our base जागर् jâgar is after all not to take Vṛiddhi, and hence a new special rule (Pâṇ. VII. 2, 5) settles the point by granting to जागृ jâgṛi a special exception from Vṛiddhi, and thereby establishing its Guṇa. No wonder that these manifold changes and chances in the formation of the First Aorist of जागृ jâgṛi should have inspired a grammarian, who celebrates thern in the following couplet:
गुणो वृद्धिगुणो वृद्धिः प्रतिषेधो विकल्पनं ।
पुनर्वृद्धिर्निषेधोऽतो यणपूर्व: प्रामयो नव ॥
"Guṇa, Vṛiddhi, Guṇa, Vṛiddhi, prohibition, option, again Vṛiddhi and then exception, these, with the change of ऋ ṛi into a semivowel in the first instance, are the nine results."
Another difficulty consists in the want of critical accuracy in the editions which we possess of Pâṇini, the Siddhânta-Kaumudî, the Laghu-Kaumudî, the Sârasvati, and Vopadeva. Far be it from me to wish to detract from the merits of native editors, like Dharaṇîdhara, Kâśînâtha, Târânâtha, still less from those of Professor Boehtlingk, who published his text and notes nearly thirty years ago, when few of us were able to read a single line of Pâṇini. But during those thirty years considerable progress has been made in unravelling the mysteries of the grammatical literature of India. The commentary of Sâyana to the Rig-veda has shown us how practically to apply the rules of Pâṇini; and the translation of the Laghu-Kaumudî by the late Dr. Ballantyne has enabled even beginners to find their way through the labyrinth of native grammar. The time has come, I believe, for new and critical editions of Pâṇini and his commentators. A few instances may suffice to show the insecurity of our ordinary editions. The commentary to Pâṇ. VII. 2,42, as well as the Sârasvatî II. 25, 1, gives the Benedictive Âtmanepada वरीषीष्ट varîshîshṭa and rate स्तरीषीष्ट starîshîshṭa; yet a reference to Pâṇ. VII. 2, 39 and 40, shows that these forms are impossible. Again, if Pâņini (VIII. 3,92) is right—and how could the Infallible be wrong?—
