to have
been lost, so that she watched the mouths of others closely when they
were talking, and labored painfully after the sounds. Finally, she
dropped her mimicry of language, and, at first very slowly, acquired
words with the ordinary infant pronunciation, showing a preference for
labials (_p_, _b_, _m_) and linguals (_t_, _d_, _n_, not _l_). Presently
she substituted easy sounds for difficult ones. In the period from
eighteen months to two years of age, the following defects of
articulation appeared regularly: _v_ was pronounced like _b_, _th_
(_this_) like _d_, _th_ (_thin_) like _t_, _z_ like _d_, _s_ like _t_,
_r_ like _w_, _j_ like _d_, _ch_ like _t_, _sh_ like _t_; further:

Initial. Final.
_f_ like _w_, _f_ like _p_,
_l_ not at all, _l_ correctly,
_g_ like _d_, _g_ correctly,
_k_ like _t_, _k_ correctly,

and in general correctly, _m_, _b_, _p_, _n_, _d_, _t_, _h_, _ng_, _w_.
On the other hand, the initial sounds _bl_, _br_, _li_, _pr_, _fl_,
_fr_, _dr_, _tr_, _thr_, _sp_, _st_, became _b_, _b_, _p_, _p_, _w_,
_w_, _d_, _t_, _t_, _p_, _t_; and the initial sounds _sk_, _sw_, _sm_,
_sn_, _sl_, _gl_, _gr_, _kw_, _kl_, _kr_, _hw_, became _t_, _w_, _m_,
_n_, _t_ (for _s_), _d_, _w_, _w_, _t_, _w_, _hw_ (_h_ weak). The letter
_y_ was not pronounced at all, at first.

From this table, as Humphreys rightly observes, may be drawn the
following conclusions in regard to the initial sounds of words:

When a letter which could be pronounced correctly preceded another, the
first was retained, but, if both were represented by substitutes, the
second was retained. If, however, the second was one which the child
made silent, then she pronounced the first. Thus, _tr_ = _t_, _kr_ = _w_
(for _r_), _kl_ = _t_ (for _k_, _l_ being one of her silent letters).
With these results should be compared those presented in regard to
German children, in the paper of Fritz Schultze (p. 239 above) (which
likewise are not of universal application).

The accent was for the most part placed on the last syllable. Only one
case of the invention of a new word could be established. When the child
was about eighteen months old, a fly flew all about her plate when she
was eating, and she exclaimed, "The old fly went wiggely-waggely." But
at this time the child had already learned to speak; she knew,
therefore, that perceptions are expressed by words. Notwithstanding, the
original invention remai