Secret Origins of Dog Food


When did commercial dog food come to be? When did the family pooches get the upgrade from bones and potato peelings to delicious canned horse meat? This question arose when I wanted a character to bring up* dog food in “From the Earth to the Zoo.” But I came to wonder if dog food even existed in the mid-1930s, when the story is set.

Most of what I learned comes from this article at Neatorama, which is a reprint of a 1988 article in Uncle John’s Unstoppable Bathroom Reader. I challenge you to find a more reputable source.

The short answer is: Yes. Dog biscuits came first, in 1860. According to the story, an electrician named James Spratt saw some dogs in the port of London greedily munching up old ship’s biscuit, a mixture of flour, water and salt that was a staple for sailors during their long sea voyages. Ship’s biscuit was also called hard tack, which tells you just how appetizing it is. Spratt added vegetables and beef blood to the recipe, and thus the first dog food was healthier and probably more appetizing than what the poor sailors had to choke down.

Canned dog food came in 1922 in the form of Ken-L-Ration, and was primarily horse meat. The story of pet food’s growth in popularity is a story of advertising creating a need. In the 1960s, pet food industry lobbying groups (have you ever met a pet food lobbyist? Me neither.) cooked up a campaign to convince consumers that feeding dogs anything except commercial dog food was a bad idea. Their message: Table scraps kill.

*Bring up in conversation, not upchuck.

Did they say that in the 1930s?

My love affair with Google’s Ngram viewer began when I was drafting “The Apprentice’s Magician” and threw down the phrase “exit strategy.” Immediately I stopped, doubting that the phrase was used in the 1930s. But how to find out?

Happily there’s a tool that helps: Google’s Ngram Viewer. The viewer generates graphs of the frequency of words in the texts that comprise Google Books over time. So punch in “pizza” and here’s what you get:


The viewer accounts for the fact that many more books have been published in recent decades, and normalizes the results by the number of books published each year. Which means the results reflect how popular or not a word or phrase was at a certain time, not simply how often it was printed.

But what about more everyday terms? In the mid-1930s was it more common to say pants or trousers, for example? I feel like trousers might have been the thing, but I don’t know. The Ngram viewer can chart multiple words or phrases together for comparison. Because the Mabel Brightwing series is set in the mid-1930s in New York, I narrowed the search parameters to 1900-1950, and only American English.


Ah ha, throughout the early 20th Century, trousers led pants. There’s an interesting dip in both terms starting around 1945. Why? This is where I get pulled down the rabbit hole. Maybe jeans started to take over? Let’s add that, pull out to the whole century, and see what we get.


Well, there’s still no obvious explanation for the mysterious reluctance to talk about pants, trousers, or leg wear in any form from the late ’40s to the late ’60s, but we can see that right around 1969 pants took over from trousers, with jeans also on the rise.

The viewer can also chart phrases (and a lot more—see the About page for all the mathy-lingustic details). At one point in “The Flight of the TKTK” I had a character say “four-legged friend,” but I became seized with uncertainty. Did Americans say that in 1936? How much?


Turns out they did, but very rarely. I noodled around for alternatives, and hit upon the sexist “man’s best friend.” Let’s take a look:


Yup, that unfortunate phrase appeared about ten times more often in the early 20th, and charged up even more starting around 1935. I went with it because the character who utters it is a product of his times, and an unenlightened product at that.

So back to “exit strategy,” were my instincts right? They certainly were:


It strikes me that a talented programmer could use the Ngram viewer’s database to create an Anachronism Detector. Feed it a piece of historical fiction, give it a year or date range, and the program could chew through the document and highlight words and phrases that don’t belong. Seems like a good project for some Googler’s 20% time, although that itself may be a myth. This article sheds a little light there.