But she didn't remove vellus hairs from the equation! "Hair weight" refers to the total weight of grown hair; that is, the total of vellus, intermediate, and terminal hairs. The vellus hairs in a typical balding guy can obviously be a rather significant fraction of that total weight.

Here's what _I_ think is the difference in the hair-growing effects of Propecia and Rogaine: if a given treatment has very little increase in total hair-counts at the same time that it significantly increases the total weight of regrown hair (which is close to what Propecia does), then it seems clear that its main effect is to increase the thickness of the hair that's already there; in fact, some of the pre-existing vellus hairs probably become intermediate hairs, some of the pre-existing intermediate hairs become terminal hairs, and some of the existing terminal hairs become even thicker than they were. But if another treatment produces an increase in hair-counts which is also almost numerically identical to its increase in total hair-weight (which is close to what Dr. Price found in her minoxidil study), then that suggests that maybe ALL hairs are increasing in rough proportion to that overall number. That is, brand-new vellus hairs are being born; some of the old vellus hairs become newer intermediate hairs; some of the old intermediate hairs become newer terminal hairs; and some of the older terminal hairs become even thicker. And that idea is supported by older studies which showed relative increases in ALL those classes of hair from minoxidil treatment.

Again, Price et al counted and measured the weights of ALL hairs. Here's a brief description of the procedure from the recent Propecia study, which is also identical to what they did in the previous minoxidil study:

"Evaluation procedures. Marking and hair clipping method. At the first clipping (week -12), a plastic template with a square hole (1.34 cm^2) was placed over the chosen scalp site. All hairs within the designated area were pulled through the template with a crochet hook, grasped, and carefully hand clipped intact with straight surgical scissors under a magnifying light, which left the hair approximately 1 mm in length at the scalp. After removal of the template, 2 small, permanent tattoos were placed in nonadjacent corners of the square clipped area with the use of a Spalding and Rodgers marking apparatus. [...] At each subsequent 6-week interval visit, the plastic template was placed over the target area; the 2 small tattoos were used for accurate placement. The hair in the target area was carefully hand clipped as previously described. The clipped hairs were placed on prefolded, prelabeled collection paper, the paper was folded to form a packet, and the packet was placed in a small glassine envelope and stored until the hair samples were weighed and counted at the end of the study.

"Hair weight determination. At the end of the initial 48-week study (week 48), hair samples collected at weeks -6, 0, 6, 12, 24, 36, and 48 were weighed in a single session by a technician who was blinded to treatment, subject, and visit number (ie, time). [...] Before hair was weighed, each clipped hair sample was degreased with hexane, dried, and conditioned for at least 24 hours in a room with a constant temperature and humidity (20 degrees C, 65% relative humidity). The hairs were then transferred to a weighing pan and weighed in a computerized balance with a precision of 0.01 mg. After hair was weighed, the hairs were quantitatively returned to the packet.

"Manual hair counts. After the hair samples were weighed, the individual hairs from the samples were laid out on a marked grid in groups of 5 and counted manually. This procedure was carried out at the end of the initial 48-week study for hair samples collected at weeks 0, 24, and 48.... Hairs that were shorter than 2 mm in length were excluded from the count, because these could represent hair shaft fragments rather than intact hairs."

Bryan