With additional knowledge about biogeography, species ranges, species sampling coverage / description rates, already proven adult-larvae identity and so forth,

With additional knowledge about biogeography, species ranges, species sampling coverage / description rates, already proven adult-larvae identity and so forth, they can get you sufficiently in the ballpark to make a strong prediction that in popular reportage is forgiveably rendered as "sure". Wagner is a career lepidopterist and an authority on caterpillars. He will be intimately familiar with caterpillars in the family Notodontidae (which contains the genus Ursia) and know their characteristic morphology and diversity. There are only two species known in the genus Ursia. The other one has a more southwesterly range and is not known to occur in Texas. Also, the caterpillars of species in closely related genera in the Notodontidae are well-surveyed. So, looking at a shrub in the range of Ursia furtiva and finding a caterpillar that fits the bill in terms of expected morphological characters, and being able to rule out its belonging to closely related genera, it's a reasonable leap of logic to attribute the larva to that particular species only previously known by its adult form. It's not "certain" at that point, but as a biologist who spends his life researching those taxa, and knowing well where the missing gaps are, it's the kind of hunch that is very often proved correct.