Oh, wait. It's not a space alien. It's a Brazilian Treehopper.
And yes, it's the real mccoy:
But the site Why Evolution is True rather spoils things by resorting to the veriest "just so" stories rather than pursue the route of the Old Physicists: abstracting from sensory impressions (quiae) to a rational principle (propter hoc) and then deducing consequences from that principle and falsifying alternative principles to establish a physical cause. We live in the publish/perish age of Jumping To Conclusions. Gotta rack up them pubs!
The alternative is the modern Pythagorean approach of creating a mathematical model and forgetting about physical causes. But evolution is not math-friendly. There are no Darwin Equations.
In either case, hand-waving is not science:
A first guess is that it’s a sexually-selected trait, but those are often limited to males, and these creatures (and the ones below) show the ornaments in both sexes. Kemp hypothesizes—and this seems quite reasonable—that “the hollow globes, like the remarkable excrescences exhibited by other treehoppers, probably deter predators.” It would be hard to grab, much less chow down on, a beast with all those spines and excrescences.
Note, though, that the ornament sports many bristles. If these are sensory bristles, and not just deterrents to predation or irritating spines, then the ornament may have an unknown tactile function.
But notice how difficult it is to talk about evolution without talking in teleological terms. The structure is "for" something: sexual attraction (hubba hubba), or predatorial deterrence, or sensation, or something.
Of course, this was the very thing that led the atheist philosopher Jerry Fodor to rip natural selection as too inherently theological.