336 F. S. Growse —The 'Etymology of Local Names in N India. [No. 4, 
stroyed the giant Madhu. On the site of the captured stronghold the hero 
is said to have built a city, called indiscriminately in Sanskrit literature 
Mathura or Madlm-puri: the fact, no doubt, being that Mathura was origi¬ 
nally the name of the country, with Madhu-puri for its capital. In course 
of time the capital, like most Indian cities, gradually shifted its site, probably 
in order to follow the receding river ; while Madhu-puri itself, fixed by the 
locality of the wood that formed its centre, became first a suburb and finally 
an entirely distinct village. Simultaneously with these changes, the name 
of the country at large was attached par excellence to its chief city, and 
Madhu-puri in its obscurity became a prey to phonetic decay and was cor¬ 
rupted into Maholi. The transition is a simple one ; the Ji being substitut¬ 
ed for dh by the rule II. 27 Kha-gha-tha-dha-hham Hah , which gives us the 
Hindi hahira for the Sanskrit hadhira, ‘ deaf and lahu for vadhu , ‘ a female 
relation.’ 
It will be observed that Madhu-puri as a literary synonym for Ma¬ 
thura remains unchanged, and is transformed into Maholi only as the name 
of an insignificant village. Thus an easy solution is found for the difficulty 
raised by the same critic I have before quoted, who objects “ If it is pos¬ 
sible in the lapse of time to elide they) of puri, why have not the oldest 
towns in India like Ilastina-pur yielded to the change P and in the case of 
more modern towns why do we not find the change half-effected, some mid¬ 
dle place in the transition stage ?” To the former of these two questions 
I reply that a name when once petrified in literature is preserved from 
colloquial detrition. Thus, of two places originally named alike, one may 
retain the genuine Sanskrit form, while the other becomes Prakritized, ac¬ 
cording to their celebrity or otherwise. A parallel is afforded by the names 
of many English families : the elder branches retain the old spelling, how¬ 
ever much at variance with modern pronunciation, as for instance, Berkeley 
and Marjoribanks; while the obscurer branches, who seldom had occasion 
to attach their signatures to any document, conform their spelling to the 
sound and appear in writing as Barkly and Marchbanks. Or not unfre- 
quently they retain the old form, but pronounce the word not in the old- 
fashioned way but according to the value of the vowels in ordinary modem 
pronunciation. Thus Hastinapur exists unchanged, by virtue of its histori¬ 
cal fame ; had it been an obscure village it would probably have been cor¬ 
rupted into Hathaura. In fine, it may be accepted as a general rule that 
when the termination pur, pura, or puri, is found in full, the place is either 
comparatively modern, or if ancient is a place of pre-eminent note. The 
one exception to the rule is afforded by names in which the first element of 
the compound is a Persian or Arabic word. Some of them may be much 
older and yet not more distinguished than many of pure Hindu descent 
from which the p has disappeared ; but the explanation lies in the natural 
