Here's an apparently simple morphological question: why doesn't 'dinosaur' end with -us?

Compare:
dinosaur
brachiosaurus, stegasaurus, apatosaurus, tyrannosaurus [rex]

Etymonline isn't too helpful in this case:
http://etymonline.com/index.php?term=dinosaur

I can assume that "dinosaur" is just a change (it's the exception), with a loss of that final syllable.

But I also wonder: -r was a common nominative masculine singular ending in Latin, so that could explain it too. Then why the -us vs Ø?

When I started thinking about the meaning, something very interesting occurred to me:
-us seems to refer to specific species of dinosaur.
Ø is used more with generic, maybe casual usage.

So "tyrannosaurus" is a specific species (because it's the scientific name?), while a Tyrannosaur is any large flesh-eating lizard creature-- the class of all dinosaurs that resemble T-Rex's.

Likewise, a "dinosaurus" feels like a very specific kind of creature (whatever kind that might be).

This seems to hold for new coinings as well:
"Herbasaur" is a perfectly valid (hypothetical) name for a dinosaur that only eats plants.
"Herbasaurus" is a perfectly valid (hypothetical) name for a certain species of dinosaur that happens to eat only plants.


Does anyone else share my intuition on this one? Am I imagining things?
And do you have any general insight into the distribution of that -us?

Does it have to do with the -us being on single species and as dinosaur is the 'parent' grouping, it doesn't perhaps?

I have heard them referred to as tyrannosaurs, stegosaurs, etc though... hmm I'm not sure I understand why.

Yeah, I get the feeling that it's different if I said,

"a tyrannosaurus..." (meaning a specific one in this species)
"a tyrannosaur..." (meaning one of any type of this species)

My instinct says it's species vs broader type - what do you think?