For full-resolution image, click this link...then click the image again when it loads in your browser window.

The male tarantula hawk seen above was visiting flowers of the milkweed Ascelpia erosa...as were the female; male-female pair; and larger group in the posts below...all photographed at the same time and place:

         

My best attempt at the species ID for these wasps is Pepsis pallidolimbata smithi. But this may be P. thisbe. For more details on this putative ID, see remarks here.

Note that the relatively straight (actually slightly-arched) antennae in the top photo immediately suggest a male. That's because in female Pepsis the terminal portion of the antennae is usually held curled by 3/4 of a turn (or more) into a comparatively tight circular-spiral...as seen in the first two thumbnails above. However, females are capable of straightening their antennae (and often do...e.g. see photos here); and males can "arch" their antennae more than seen above...though not so tightly that the tips ever curl by an amount more than about a semicircle (presumably due to the thicker flagellomeres in males, and the shorter membranous sutures between them). So the "straight vs. curled" character for determining gender is not fool-proof (some would say it's not reliable at all...though often it works, especially when the antenna-tips are tightly curled for ≥ 3/4 of a circle in which case you almost certainly have a female).

Thankfully, there are two other better (i.e. more dependable) methods for determining gender in Pepsis (beyond, of course, examining genitalia which may necessitate a specimen in hand and/or dissection)...and they also work for most aculeate hymenoptera, though they require having a sufficiently unobstructed, high-resolution view of the antennae and/or metasoma ("abdomen"):

1) The first method involves getting an accurate count of the antennal segments. Females have 12 antennal segments (referred to, starting from the base, as the scape, the pedicel, and the 10 flagellomeres) and males have 13 (i.e. scape, pedicel, and 11 flagellomeres)...except for the anomalous case of species P. grossa, where males have only 10 flagellomeres.
2) The second method involves getting an accurate count of the number of (non-retractable) abdominal segments: females have 6; males have 7.
Point of Interest: Note that a "straight vs. curled" character for determining gender is also often cited in the hymenopteran family Vespidae...but there the roles are reversed! That is the (tips) of the antennae in males are often coiled or "hooked", whereas in females they are not (e.g. see this CJAI reference).