Where do dogs get their Vitamin C from, when they don't eat fruits?

So, although you probably never thought about it, now that I mention it, yeah, that's a good point. Where do dogs get their vitamin C from?

Well, actually, we're the screwed up ones, not the dogs. You see, almost all animals have a gene for making vitamin C. Humans do not. This is because somewhere along the long and glorious line of evolution of man, this gene became a pseudogene. This means that because of some mutation, the gene didn't code for the protein that makes vitamin C anymore. So now we have a gene that looks very much like the gene for making vitamin C in our DNA, only ours doesn't work. Although this is not an evolutionary advantage (obviously someone who makes vitamin C is more likely to survive than someone who needs to import it), somehow the gene was nevertheless removed from the species.

Most animals, including dogs and cats, can manufacture vitamin C, but primates, including humans, cannot. This is why we must eat oranges, and other fruits and vegetables which contain vitamin C, to survive. The gene has disappeared along several other evolutionary lines as well, so, for example, guinea-pigs cannot manufacture vitamin C either, and must obtain it from outside sources.

alex.tan makes a valid point, which should be taken into consideration. It's important to state, though, that

1. yes, someone who makes vitamin C is more likely to survive than someone who needs to import it. That sentence contains an implicit assumption that everything else is constant. Of course, maybe a modified gene will do other things, but that does not contradict the sentence. It's possible, like in the case of sickle-cell anaemia, that the "faulty" gene also has some advantage (in that case, some malaria-resistance), however sickle-celled people still have less of a chance to survive than "normal" people, in "regular" conditions. (In areas infected with malaria, they have an advantage in malaria resistance).
   
2. As far as i know (someone correct me if I'm wrong), the gene for vitamin C is now a pseudogene, which means it doesn't produce another protein. It doesn't code for anything anymore. So it probably doesn't serve an evolutionary advantage. Statistically, "bad" mutations get fixed in a species sometimes too.

The vitamin C gene (genes, as binary 666 insists on correcting) was not the major part of my writeup, it was more background information, which I didn't want to get into at the time, but as alex.tan brought it up, I guess I did. More information can't be too bad.

You cannot categorily state that someone who makes vitamin C is more likely to survive than someone who needs to eat it.

Genes work in weird and wonderful ways. It would not be inconceivable that the mutation in the gene that led it to no longer produce vitamin C led to it producing some other protein that, in fact, produced some greater evolutionary advantage that we do not yet fully understand. Well, it can't be that bad if we've survived this far, right?

There are two steps in the pathway to produce Vitamin C that are damaged in humans. The first is universally prevalent. The second error is known to be totally eliminated in some populations.

The error is in the gene for L-gulono-gamma-lactone oxidase. This is the last enzyme in the ascorbic acid synthesis pathway. The reaction L-GLO catalyzes (production of 2-keto-gulono-gamma-lactone) will also occur spontaneously on its own, so humans do in fact produce small amounts (15-20mg per day) of vitamin C. L-GLO is pseudogenetic in humans- no alternative product occurs.

The second error deals with lactonase. Lactonase is the enzyme preceding L-GLO in the chain. If functional it would dehydrate and cyclise L-glulonic acid to produce L-gulono-gamma-lactone. Mutants of this second type produce no ascorbic acid and must get it from external sources lest they develop a health threatening deficiency.
Quite predictably, the prevalence of lactonase dysfunction can be highly correlated with the availability of vitamin C. In certain tribes of desert nomads the dysfunction is totally absent. It has been predicted that historical seafaring communities would have also had a higher percentage of functional lactonase enzymes due to the disproportionate loss of sailors with the malfunctioning allele because of scurvy.