Update: Since we posted this recipe, we’ve learned that our fish custard might have been tastier had we prepared it using different methods and ingredients. Please see the comments for a variety of helpful suggestions. And if you successfully recreate this dish, please let us know!
Some recipes should stay in the archives.
We’ve had surprising success so far with these early modern recipes. All have been edible, most have been pretty tasty, and a few – like the inaugural mac and cheese and some spiced “jumball” cookies we’ll tell you about soon – have been downright great. So, we thought, let’s branch out and have a more daring culinary adventure. When Marissa found this recipe for “fish custard” (that’s right, Marissa, I’m blaming you), we thought immediately of Doctor Who’s infamous snack:
The Doctor makes fish fingers and custard look pretty tasty. And while we suspected that fish custard might not prove our favorite recipe from this project, how bad could it really be?
Bad. So. Very. Bad.
This fish custard comes from UPenn Manuscript LJS 165, a collection of recipes in multiple hands, written and gathered together sometime between 1690 and 1802. Readers could consult the collection to find other culinary recipes but also to find out about various household remedies, like how to cure colic (presumably, by not making someone eat this dish) or to kill moths (probably by setting out a bowl of fish custard, thereby driving all living things out of the vicinity).
Please don’t try this at home. No, really. Please don’t.
One pound of Almons beat them small, in the beating
put in the Row of a Pike 4 dates cut and the yolkes of
4 Eggs temper it with cold water Straine it through a
Strainer & make a quart of it Season it with Suger Rosewater
Salt pxxxxe beaten Mace When it is Baked scrape suger on
1 c. ground almonds
1 to 1 1/2 tbsp. fish roe (ours was salmon)
3 dates, seeded and roughly chopped
2 eggs + 1 egg yolk
1/4 c. whole milk
1/4 c. sugar
1 tsp. rosewater
1/8 tsp. ground mace
a few pinches of salt
Preheat the oven to 350F. Butter a small casserole dish. Stir together all ingredients, then spread evenly in casserole dish. Bake for 15-20 minutes. Remove from oven and cool for at least 10 minutes before serving.
We quickly realized that this wasn’t going to be a traditional custard – the ratio of almonds to dairy is much too high to produce anything like a creamy texture. (The original recipe did include straining, but that would have removed all of the almonds and dates, which seemed counterproductive.) To make the mixture stir-able, we added a few spoonfuls of milk. We were unsure how vigorously to beat in the fish roe: should the eggs be broken down and, liquified, dispersed evenly throughout the custard? Or should they maintain their shape? We erred on the side of folding them in gently. It’s possible that we should have put the whole mixture in the food processor; this might have improved the final texture somewhat, though it’s unlikely to have helped the taste.
This recipe raised some interesting questions for us about interpreting early modern culinary instructions: with other recipes, we’ve had some idea of how they would turn out, especially when we started cooking and realized that they resembled some modern-day counterpart. This similarity provided some guidance; even when the original recipe’s instructions weren’t quite clear to us, we could extrapolate from other knowledge and proceed with some degree of confidence. The addition of the fish roe, in fact, threw us off less than the realization that this “custard” would not resemble anything we would call by that name. We were apprehensive – which seems a valid reaction to a fishy dessert – but also curious. What would the texture be like? Would the fish roe somehow pair beautifully with the almonds and dates in a salty-earthy-sweet combination?
Big surprise: dates, almonds, and fish roe don’t play well together. And the rosewater just made things worse. The “custard” resembled a bar cookie: very firm and sliceable into squares. In fact, it was quite dry, to the point that even if it had tasted good (ha!), eating more than a few bites wouldn’t have been very appealing. Another texture issue: baked fish roe either explodes warmly when chewed or takes on an off-putting rubberiness. We took tiny servings and managed a spoonful. (I think the fact that we did so speaks highly to our research initiative.)
So, was this failure our fault? The fault of the recipe? Should we write off early modern palates as utterly mystifying? Was the mere existence of this recipe a joke from the time-traveling Doctor? We’re willing to believe that the original execution of this recipe was probably more appealing than our effort, though we doubt that this would ever have tasted good.
Readers, we did this for you. You’re welcome. Now, please excuse me while I go brush my teeth again.