From my reading tonight:

"I have allowed (QUANTITY) to enter these relationships nonlinearly by introducing a quandratic in quanity (QUANTITY-SQUARED) in (1) and using the natural logrithm of quantity (LOG-QUANTITY) in equations (2) and (3)."

and

"We can obtain estimates with desirable asymptotic properties by specifying the likelihood function of the sample, given the nature of the sampling truncation and then solve for the maximum likelihood estimates (MLE) of the coefficients of interest."

You just know that this author never gets invited to any of the good parties.