Category Archives: recipes

Hands-on cooking and the occasional comment about recipe-writing

Vanilla slices – glorious, perhaps mundane, quite mysterious

I’ve been thinking a lot about the good old vanilla slice recently. It’s one of those standard, humble Australian bakery items that qualifies as “good old” for its longevity and as “humble” because of its modest level of finesse. But there are mysteries in all this (and questions for readers at the end).

A little journey

I grew up eating vanilla slices as my preferred after-school snack, albeit constrained somewhat by the disappointing options opposite my high school. I could choose between the milkbar’s revolting, rubbery Four’n’Twenty vanilla slices (truly deserving of the colloquial name “snot block”), or the hot bread shop’s Greekified version, filled with a semi-translucent lightly lemony paste. Whenever possible, I’d find my way to the Ferguson Plarre bakery, as their vanilla slices were the childhood paragon of gooeyness.

What has brought me to this little discourse was a rare daytrip to the countryside: Woodend, a small town an hour’s drive from Melbourne, has somehow become famous for its vanilla slices. Two bakeries vie for attention. The smaller of the two, Woodend Bakery Café, supplied a not-quite-fresh slice (the pastry was too close to cardboard, despite its very promising appearance) with a creamy sweet custard and clear vanilla notes. A bit tall (perhaps aspiring to life as a millefeuille), it was a greater-than-usual challenge to eat this vanilla slice gracefully (perhaps a mutually incompatible concept anyway).

The second bakery, Bourkies Bakehouse, offers tray upon tray of slices, some with icing flavoured with passionfruit or raspberry. These were fresh (the turnover is enormous), with a custard of similar characteristics to their competitor and stodgy pastry, but of slightly more jaw-friendly dimensions.

While I preferred the texture and dimensions of the latter vanilla slice, both co-eater Mittens and I found the custard of the former slightly tastier. It was a very close call, though, and likely to change depending on freshness and minor variations in formulation. The review on the apparently now dormant Vanilla Slice Blog was scathing of the Bourkie’s slice and praising of the Woodend Bakery Café, but on our visit the differences simply weren’t that marked.

Back home in Melbourne, we tried my childhood favourite from a Ferguson Plarre outlet. This a no-nonsense affair: (1) thick, stiff but airy custard, (2) moderately thick, unimpressive pale pastry, (3) moderately thick, sticky icing. Texturally, I think the custard is great (for a vanilla slice) – my preference is definitely for custards that are less creamy than those in Woodend. The flavour, however, was that of cream and just the faintest hint of vanilla. I like eating them, but on closer inspection, there’s not a lot that’s truly impressive. Somehow, it just all comes together to be something innocuously pleasant. Maybe life just has to be that way sometimes!

This article was originally going to include some baking in my kitchen, but I got sidetracked by some interesting observations, started by the fact that the Wikipedia entry for Millefeuilles dumps a whole pile of things together rather clumsily and imprecisely. (For a beautiful homemade millefeuille, see Sarah’s post.)

So where are the vanilla slices at home?

Although most of the people I know take the vanilla slice for granted as an Australian great, most are unaware that it not only has other names around Australia, but actually occurs in some form in other countries. I’m not talking about the French millefeuille, the distinctly classier relative, nor the Australian “French vanilla slice” which is more of a nod to the millefeuille, sometimes a mix of layers of custard and whipped cream, perhaps with coffee icing. I read once that New Zealanders and USAmericans call vanilla slices “napoleons”, but internet searches favour either clumsy millefeuilles or a layered slice of cake, custard and puff pastry under that name. Sometimes there’s even jam involved! It seems that “custard slice” or “custard square” might be a much closer relative of our target.

The state of Victoria seems to be disproportionately represented in online searches for vanilla slices. Even taking into account other names (it seems (some) Queenslanders prefer “custard square”), Victoria seems to be in the lead. There’s some bias because of the annual competition in the Victorian town of Ouyen.

Of further interest, there are remarkably few recipes for vanilla slices in common older Australian cookbooks, which makes me wonder if the popularity is less national than regional. Cooking queen Margaret Fulton, for instance, has no recipe in her main book or encyclopedia. Some Australian Women’s Weekly books do feature a recipe. The recent NSW Country Women’s Association Cookbook doesn’t. It’s not unusual for there to be no common recipes for baked goods that were only produced commercially, but it does make it harder to trace the history of the thing!

The majority of (more recent) recipes combine custard powder with milk or cream and egg. Most overseas recipes seem to omit cream. (You also have to ignore the “blinged up” recipes that have aspirations to be a millefeuille, so leave the recipes of Bill Granger and Maggie Beer and countless others out of the equation.)

And what should a vanilla slice be?

For me, the prototype of a vanilla slice is a square about 7cm (?2.5″) wide and about 4cm (?1.5″) high, with soft (not runny) white icing. The stiff custard is not a rich crème pâtissière and has a slightly aerated character (see pics). It is not lightened with whipped cream. The pastry is never particularly crisp (alas).

For my elderly neighbours, the size they remember from their childhood (the 1950s) is as above, as is the icing, but they remember the custard as being very pale and not as rich (perhaps a lean milk custard set with flour or gelatine?)

And my mother remembers a firm, pale yellow custard in Sydney’s vanilla slices. Her opinion of vanilla slices in Brisbane in the late 1960s is not flattering: “… the custard was nasty in that it was thick and solid but didn’t taste right for a custard and seemed to be mostly cornflour. In retrospect, maybe they used a lot of water rather than all milk to make the filling.”

Neither my neighbours or my mother can find a recipe from the time, so vanilla slices might well have been the preserve of the commercial baker. (I can find an English recipe for “vanilla slices” from the 1950s, which features white icing on top, and consisting of three rectangular layers of puff pastry, sandwiching two modest layers of both jam and either whipped cream or “confectioner’s custard”.)

I think the culinary world needs clarity, so tell me, dear readers (especially Gen X and older), did you grow up with the Victorian-style vanilla slice in your neck of the woods. Or was there something with a heavier custard? What was it called? And how long ago? Or was there nothing? And what about overseasy people?

My apologies in advance for any rumbling stomachs these questions may cause.

Mint ice-cream, and ice-cream machines

Have you ever met one of those people who looks ascance at you when you suggest having an ice-cream in winter? “But it’s 8 degrees and raining!” they exclaim. So bloody what? Ice-cream is can be the stuff that dreams are made of, so of courrrrrrrrrrse you should eat it all year round.

When I first visited Germany it was mid-winter and verrrrrrry blllloody ccccolllld. Didn’t stop me buying a Cornetto, but! (The shop owner almost did, as I learned abruptly that customers d o  n o t get to open the ice-cream cabinet for themselves.)

So here we are at the winter solstice (in the southern hemisphere). As I haven’t tried enough recipes from the ice-cream book I reviewed a few months back (Ice Creams, Sorbets and Gelati: The Definitive Guide), and because KitchenwareDirect were nice enough to send me a new ice-cream maker to try out when they saw that I’d lost the paddle for my old one, I decided to do a flavour that straddles ideas of cold and wishes for warmer weather: mint.

Mint in any dessert can range from sophisticated and warming to sickeningly sweet and cloying. Anybody who pairs milk chocolate with mint, for instance, probably deserves very painful punishment, especially if the mint is communicated through little flavoured pieces of candy in the chocolate. Blurk. Ick. Pfahhg.

A dark chocolate is the way to go, but I don’t think you need to prefer the darkest on the spectrum – a good dark chocolate in the 50-70% range is appropriate for most applications. If you use fresh or dried mint, things get especially interesting and complex. One of the most surprising macaron flavours I did last year for Café Liaison used dried peppermint in a dark chocolate ganache. In the ice-cream here, I used fresh common mint straight from the garden.

The machine I’m now using (on the left) is a Cuisinart Ice Cream, Yoghurt & Sorbet Maker, 2-litre capacity. It’s stylish stainless steel looks aren’t quite matched by performance, as I found a full 2-litre batch didn’t freeze well enough on a warm autumn day, but it did well for a 1-litre batch in winter. (And let’s face it, probably all ice-cream maker ice-cream still kicks store-bought ice-cream out of the water!)

The mint ice-cream from the abovementioned book combines a fairly standard French-style egg custard with some sprigs of mint. The mint steeps in the cooling custard for an extended period, after which the mint is squeezed and the custard strained, the cream added, and the mixture churned. The authors’ recipe for 800ml of ice-cream used four 10cm sprigs. I upped that to five and it could possibly have been taken further.

The Cuisinart uses a rotating frozen bowl with a fixed paddle, rather than a rotating paddle inserted in a frozen bowl (as in my older Krups model, also pictured). The Cuisinart design – now quite common in ice-cream machines – seems to make for a noisier experience. I wouldn’t have wanted to be stuck in the kitchen while it churned a full mixture! Alas, the quite clever paddle (you can see it directing the flow in, out and up) is an absolute bugger when it’s time to decant the ice-cream into a container – too much surface area picking up the fresh ice-cream as you withdraw the paddle from the container, and many angled and bumpy surfaces to try to get that ice-cream off. I also found that it didn’t scrape the sides and base effectively enough while churning, leaving a 2mm layer of hard-frozen butterfat and custard (and no hard spatula is supplied to scrape it off).

The outcome was a smooth, wonderfully delicate mint ice-cream, lacking any giveaway green tinge. I had intended serving it with a chocolate sauce, but it would have been overpowering. A dusting of powdered cocoa might have done the job instead!

The next mint escapade will be with peppermint (once I’ve grown some!).

Notes about choosing an ice-cream maker: how well a machine performs is impossible to tell just by looking at it. I strongly recommend reading reviews or asking among your friends – even fancy, expensive models with their own compressor aren’t always reliable. As long as you don’t expect to make ice-cream without forethought, the common and fairly affordable machines with bowls that need to be frozen for 24 hours before making the ice-cream are a good way to go, but I do recommend making less-than-maximum sized batches in order to reduce unhappy sloppies. It’s also advisable to chill the mixture in the fridge for many hours beforehand.

In the case of the Cuisinart, the customer reviews at KitchenwareDirect are strongly positive (aside from the machine noise), but I am less positive – it does the job, but could be better. Both my now-paddleless Krups model (GVS2: highly rated on Amazon Germany) and my Mum’s 1985 Phillip’s machine (still going strong) are more effective at cleaning the sides and slightly quieter, but as Krups appliances are no longer readily available in Australia… Sniff.

My thanks to Brad at KitchenwareDirect for providing me with a new ice-cream machine.

Fantastic chocolate ice cream, plus equipment failure

I can feel myself approaching an equipment crisis. My favourite food processor is no longer sold in Australia and spare parts are crazy expensive. My kitchen scales are on the way out. My new spatulas bend too much. And just recently I lost the paddle for my ice cream maker. All such things are discovered at the least convenient moments, and the last two had to happen just as I was reviewing a book on ice cream – Ice Creams, Sorbets & Gelati by Caroline and Robin Weir.

Some weeks after testing a few recipes, I noticed the book had a recipe that ostensibly reproduced the stunning cacao extra bitter ice cream at Berthillon in Paris (see pic here). As that was the most impressive chocolate ice cream I’ve ever eaten, I had to try it at home.

I made the custard. Chilled it for 24 hours. Made sure the ice cream maker’s bowl was in the freezer.

Took out the custard. Assembled the machine. Couldn’t find the paddle. Couldn’t find it. Anywhere.

I put the custard back in the fridge. The bowl back in the freezer. And searched.

I went online to see if I was misremembering the colour or shape of the paddle while rummaging through cupboards and drawers and down the back of shelves.

In the end, I had a pot of custard and a frozen bowl and no paddle, so it seemed the only options was to try to churn the mixture by hand with a rigid spatula (that’s when I discovered the new spatula isn’t strong enough). Despite only having made a half mixture of custard, this was hard work, especially as you don’t want the mixture freezing solid against the walls of the bowl (that’s why the paddles scrape down the walls and base constantly). Uff.

As much as I complain, it was educational. Doable. And the result was damn good.

At first, the coffee in the mixture was too noticeable and I wasn’t excited. But 24 hours later this had mellowed and probably wouldn’t have been noticeable to new tasters. The final ice cream was dense and rich, beautifully chocolatey and melted superbly. A textural delight.

I won’t reproduce the exact recipe here, but the key characteristics were a basic rich vanilla ice cream custard with quite a lot of cocoa and chocolate dispersed through it, along with some sugar syrup and a small amount of instant coffee. (The cocoa is boiled for some minutes to get rid of the raw taste, and the chocolate is then melted into the milk for the custard.)

Was it the same as Berthillon’s? The potential was there. I used a 55% cocoa solids dark couverture. If I were to make it again, I’d try 70% or so as I think that would get it closer to the mark.

Meanwhile, I’m off to borrow my mum’s ice cream maker from 1985. She still has her paddle.

Where there’s smoke there’s crème brûlée

Is Paris burning?

No, it’s just Harry making crème brûlée again…

[My Parisian correspondent, Harry de Paris, has been wanting to write about these lovely custards for Syrup & Tang, so here you are… my first guest writer. Duncan.]

Inspired by the film Julie and Julia, I recently decided I would embark on a similar undertaking to cook all the recipes in a single cookbook. Armed with my beloved copy of Family Circle Dinner Parties (circa 1990), I determined every weekend to cook each of the recipes the book contains which, as it happens, are neatly organised into three and four-course meals.

The first meal consisted of an entrée of tomato garlic mussels, followed by veal with wine and mustard sauce for the main meal, and finally, crème brûlée as dessert. I must admit I was somewhat surprised to find a recipe for crème brûlée in an Australian cookbook from twenty years ago. I’d certainly never made it before, and had only tried it a few times in restaurants in the ten years I’ve been living here in Paris. They are delicious, really, but I’m just not that mad on desserts.

The first two courses of the meal went fine, both in the preparation and the eating. But when it came to the dessert recipe, I’m afraid I have to use the F-words: Fundamentally Flawed. I followed the instructions and doses to the letter, but the end product was a grainy, sometimes lumpy, tasteless custard with caramel on top.

If you haven’t enjoyed the pleasure of eating a crème brûlée, you should know that it is in fact supposed to be a smooth and light vanilla custard (thick, not runny, and certainly not lumpy or grainy). It is served chilled, but with a thin layer of burnt sugar on top.

Despite the initial hurdle, my resolve to pursue my endeavour was strong and I was determined to cook the following week’s dinner party meal. After all, I was on a mission. But sadly, my determination was no match for my desire to get those doggedly difficult crème brûlées right. Family Circle would have to wait!

Disappointed with the recipe I had followed, I delved into my cookbooks and browsed through a forest of websites and online cooking forums, only to discover that there are as many recipes for crème brûlée as there are budding chefs. In the multitude of recipes I found, the proportions of egg yolk to cream and the amount of sugar used varied as much as the cooking temperatures.

My research also revealed that crème brûlée is a dessert whose origins the French, English and Spanish all lay claim to. The Spanish (well, Catalonian) variety is known as crema catalana and, among other distinguishing features, it contains cornflour (unlike the French version). Burnt cream, on the other hand, originates from Cambridge, and is made mainly on the stovetop and can also include cornflour to thicken it. Finally, the French crème brûlée can be traced back as far as 1691 when the then royal chef, François Massiolet, wrote a recipe for it in his book of recipes, Nouveau cuisinier royal et bourgeois. In more recent times, the crème brûlée was popularised during the 1980s by chef Paul Bocuse, and has been a hit in French restaurants ever since.

Through much trial and error, I managed to learn that the French version of crème brûlée is best cooked in the oven at a very low temperature of around 100°C. After all, crème brûlée is a kind of custard, so if the internal temperature manages to get anywhere near that high, you’ll make it curdle. Trust me, on my second attempt, I did! You see, my French oven is powered by gas, and it has only two possible temperatures: a minimum of 160°C, and a maximum which exceeds that of a kiln.

Trying to navigate this French eccentricity, I discovered a further point of discord among recipes: to bain-marie or not to bain-marie. In my case, achieving such a low temperature could only be possible by putting the custard-filled ramequins in a tray of water. As my learned friend Duncan explained to me, the water surrounding the ramequins boils at 100°C. Any hotter and it turns into steam, so the submerged walls of the ramequins (and therefore the edges of the custard) can’t get any hotter than the temperature of the surrounding water. This turned out to be a key factor in getting the custard to set without it curdling. It takes an eternity in the oven, but it seems to work.

A further hint I found on the internet was using a small amount of egg white. This suggestion tends to make waves among purists, who claim that only the yolks should be used, but it doesn’t hurt and it also appears to help the custard to set.

Now that I had finally managed to make palatable crème, I had to “brûler” (burn) them. When making crème brûlée, it really is worth doing the burnt sugar at the last minute, just before you serve them, because the contrast between the chilled custard and the hot, crispy layer of sugar is what makes it so delicious. Nowadays in most restaurants in Paris, chefs tend to use a gas torch, and incinerate a thin layer of sugar sprinkled across the surface of the chilled custard. I had no such torch to begin with, and somewhat precariously held the custards on their side over the flame of one of my gas stove burners. Miraculously, the custard never ended up sliding out of the ramequin and onto the stovetop, but I knew that disaster was nigh and this makeshift solution would never do in the long run.

That was when I discovered the crème brûlée iron (fer à brûler or fer à caraméliser in French) – a round metal disc with a long handle on it, which you heat over the flame. When the metal is searingly hot and about to turn red, you brand the custard with the iron, burning the sugar and turning it to a delicately thin, crisp layer of caramel. This is apparently the way crèmes brûlées were originally made, and it’s loads of fun to do. The smoke generated by this method, however, is a little worrying for the neighbours, who at this point were looking across the street and into my kitchen to make sure I hadn’t started a fire!

Recipe (makes four individual crèmes brûlées)

4 egg yolks (plus a little of one egg white)
65 g white sugar
460 ml cream (30-35% fat)
Several drops of vanilla essence
Brown sugar (demerera or dry crystal brown sugar; normal sugar can also be used)

1. Preheat the oven at 100°C if possible, or at its lowest temperature if not.
2. Slowly heat the cream and the vanilla together in a saucepan until it reaches a simmer.
3. In a separate, large bowl, whisk together the egg yolks and the white sugar.
4. While the cream is still hot, gradually pour it into the egg and sugar mixture, whisking all the while.
5. Let the mixture stand so that the bubbles that formed from the whisking break. It depends how much you whisked, but leave it at least 30 min.
6. Once the mixture has settled, pour equal amounts into each crème brûlée dish.
7. Place the filled dishes in a bain-marie (use a large oven tray placed in the oven) and fill the tray with water so that the tops of the crème dishes are slightly above the level of the water.
8. Cook the crèmes in the bain-marie (in the oven) for 60-90 min, until they are just set (the middle may still seem runny). Remove them from the oven and cool on the benchtop before placing them in the refrigerator for long enough to chill.
9. While the crèmes are still cold, sprinkle with the sugar and caramelise using the heated crème brûlée iron, a blow torch, or under the grill.

Thanks for reading!

Beef short ribs cause ecstacy

I’m not a fan of small pieces of meat on the bone. Chopped up duck, chicken wings, ribs, bak kut teh… I’ve never enjoyed chewing modest amounts of meat off bones. For years I ignored a cut of beef – short ribs – believing it would have, well, ribs in it. Duh. In fact, the short rib cut is often sold boneless in Australia, leaving just luscious layers of very flavoursome beef and quite a bit of fat. This cut is the relatively thin layer of meat that covers the outside of the ribs of the animal on the side of the ribcage, beyond the fleshier back areas (with thicker bone) used for cuts such as Scotch fillet (rib-eye) or rib roast.

While I was working on my Where are the Good Meat Books? feature on The Gastronomer’s Bookshelf, I came across a recipe for Korean-style oven-browned short ribs in one of my favourite meat books, The Complete Meat Cookbook by Bruce Aidells and Denis Kelly. It sounded very tasty, so it was time to get ribbin’. The outcome was nothing short of heaven. Fatty heaven, but heaven nonetheless.

The idea is to braise rib meat with lots of garlic and ginger until tender, then roast the pieces of meat in a hot oven until they crisp a bit. I must admit that, despite making this dish four times, I’ve never done the final step. It’s just too delicious in its braised form that I didn’t bother with the final oven crisping!

You can use bone-in or boneless short ribs. Trim them of excess fat and cut into largeish mouthfuls. Pop them in a good braising pot (cast iron is great). Add lots of ginger, garlic, soy sauce and some brown sugar, spring onions (scallions) and a little vinegar or lemon juice. Add water to cover the ribs. Simmer uncovered until the meat is tender (may take two hours).

Towards the end, you’ll need to stir more frequently as the moisture evaporates and things stick a bit. The amount of fat which renders from the rib meat is considerable and means that you get that lovely rich browning that readers might know from some south-east Asian beef curries/stews. The result is a dish with a deep, aromatic savoury-sweet profile with tender chunks of meat that separates into coarse layers. Rounded off with a little sesame oil at the end of cooking, it’s fantastic.

Serve it in relatively small portions on rice. Small portions? Yes, because everyone will want seconds! I’d recommend accompanying it with clean sweet preserved vegetables (light flavours), or some fresh, lightly cooked snowpeas or buk choy or wilted spinach.

Breakfast time and pancake temptation


For the most part, I’m not the kind to dally over, under or prior to breakfast. Once upon a time, I could barely manage the few minutes of showering before my stomach felt like it was devouring me. Staying with various friends in Germany was a special torture, as the morning routine extended beyond showers and shaves, to setting tables, brewing coffee, laying out cold meats and cheese, and popping down to the bakery for fresh rolls. At times, breakfast begins more than an hour after rising. (It must be said that such German breakfasts are delightfully homely affairs — totally gemütlich — and can fill a day with bonhomie. But only if I haven’t already collapsed from the aforementioned self-consumption.)

Breakfast, as some readers already know, is rather important to me, associated with simple but very important dietary specifications. It is, however, here that I first reveal the temporal parameters of my breakfast. As a child, I would eat before showering. Little did I know this would establish digestive patterns that have turned me into a wimpering hunger-immobilised wreck when visiting Germans or poorly-catered hotels.

As a student, I tried going to the gym before breakfast, and quickly found myself staggering, zig-zag fashion, across campus as my nervous system gradually shut down non-essential functions. I think it was around about this time that I discovered morning croissants were not only delicious, but an excellent surrogate for gym exercise.

After quite a few years of travelling, I can now manage almost 45 minutes without food before my stomach starts gnawing. And you know, 45 minutes is just right for making puffy pancakes. I grew up with my mother making pikelets (the only true puffy pancake in my family) with butter and golden syrup for special breakfasts, and of course she was up well before me, so there was never a delay in product delivery. As an independent adult, these morning luxuries have been rare, because they take too long.

However, I was so attracted to the recent tasty photo of banana/bacon/ricotta/maple pancakes on The Last Appetite a few weeks back that I decided to brave the ticking clock and reproduce them at home. 45 minutes. Done, dusted, devoured.

I rarely post recipes on Syrup & Tang because most readers have their own trusted books, or you can Google pretty much anything (with about an 85% chance the result will have been “adapted” from a book). So, off you go and find some fat pancakes (often called American pancakes. I like the buttermilk pancakes in Stephanie Alexander’s big fat Cook’s Companion). Slice or mash some banana. Fry some streaky bacon over low heat for quite a long time until fairly crisp. Have heaps of butter and maple syrup ready! Try not to forget the ricotta whipped with honey (ehem, damn). Devour noisily.

And if that doesn’t suit you, have a bowl of Rice Bubbles. They’ve made me the man I am today.

Postscript: My co-eater, Mittens, was browsing through Maggie Beer’s Maggie’s Harvest and came across a salad with pear and prosciutto. “How about this on our next batch of pancakes?” What a good idea… (not the salad part, of course).

Canelés de Bordeaux (or: my love of copper moulds)


I feel myself drawn inexorably to baking projects that have no end, just a rocky path of discoveries, flops, experiments, successes, flops… Not content with macarons and pastéis de nata to add girth and grey hairs to my existence, I started playing with canelés de Bordeaux about a year ago. Why? Why? The lure of something that absolutely required expensive copper moulds and beeswax must have clinched the deal. I let on to readers that I had embarked on another of my baking odysseys back in March, but delivering the goods has taken far too long. It would have been quicker if every batch didn’t require advance preparation and considerable time in the oven. At least I’m not alone, for everyone from Clotilde at Chocolate & Zucchini to the talented Julia at Melanger has had the bug, as have many people on eGullet.


Canelés are fluted cakes made of a very simple rum-flavoured batter. Copper moulds are “greased” with beeswax and filled almost to the top. The desired result is soft, almost gooey on the inside (some people say “custardy”) and a shade of rich brown on the outside — preferences range from prettily golden to charred black, though dark brown is the traditional preference.

The name canelé has often been spelled cannelé, but was codified with one N only a few decades ago. The genuine article from the Bordeaux region is designated canelé de Bordeaux, while in other places canelé, canelé bordelais or even just bordelais (plural: canelés bordelaises) are used. (I see no reason to assign two Ns to the non-local version, and it’s possible that version is the perpetuation of a typo from somewhere.)

The frustration of these strange objects is in the moulds and the cooking. It’s easy to cook something that can be as brown as you like, right? Cool! It’s easy to grease a mould so something doesn’t stick, right? Yes. But what if the question is how much something should not stick, or how to stop your batter rising a few centimetres out of the mould, or how to avoid ugly air-pockets inside? Hmmm?

Or one could wonder why some recipes insist on freezer-cold moulds while others don’t. Why some want the canelés baked on the bottom shelf of a convection oven, or not. Why some place freezing moulds on oven-hot trays, or why some add hot milk to the other ingredients while others wait until the milk is cold… the list of contradictions goes on, and despite the wearyingly long thread on eGullet, most of the crucial parameters go unexplained. So let me summarise what I seem to have found out.

  • Copper moulds (tin-lined) are cool, expensive, and yield the “genuine” texture. 55 mm diameter moulds cost at least EUR 8.50 each (currently AUD 13) in France. I have a grand total of two! Aluminium moulds are cheaper, but not quite as good. Note that copper moulds have to be “seasoned” before the first use (see Wolfert’s article link below).
  • Silicone moulds are pretty useful and can yield an interesting, shiny and smooth-as-glass crust that I think is by no means inferior to the “genuine” article. Not all silicone moulds are equal, however. An eight-canelé sheet costs at least EUR 15 in France (about AUD 25) and you would need at least two because you can’t use all the cavities (see “Baking”, below).
  • Beeswax provides a thin coating on the metal moulds which doesn’t melt until the moulds are quite hot (unike butter or oil). It adds a pleasant taste which is part of the genuine “signature” of canelés. It can be hard to buy in smallish quantities, and you would probably have to order it from afar. (I got 500 gm sent to me from a Queensland-based company.)
  • It’s possible that strong heat from the bottom is desirable to harden the exterior quickly and brown the batter at the bottom of the mould (which will form the top of the finished product).

RECIPE: Canelés bordelaises

This recipe is an adaptation of the highly respected Paula Wolfert recipe (worth reading too!). I’ve metricated it, tweaked slightly and added some more technical explanation.


Makes 10 canelés in 55mm diameter moulds. Prepare at least one day in advance.

480 ml milk
170 gm caster sugar
30 gm butter
85 gm soft flour (or about 70 gm plain flour and 15 gm cornflour/cornstarch )
65 gm egg yolk (approx four yolks)
20 ml rum <-- I don't think you need more to achieve a beautiful aroma
pinch salt

Bring milk to just below boiling (avoid boiling it) and then remove from the heat. Add a third of a vanilla bean, sliced lengthways along one edge to release the seeds into the milk. Stir.

In a food processor, throughly mix together flour and butter, then add sugar, egg yolks, rum and salt. Remove the vanilla bean from the hot milk and gradually add the milk to the batter. Strain, and then store the finished batter in the fridge for at least 24 hours.

On the day of baking you should prepare the moulds at least an hour in advance, preferably longer:

Metal moulds: heat the moulds in the oven. While doing this, place a small glass jar or a cupped piece of aluminium foil in about 2 cm water in a saucepan. Place some beeswax and a little neutral-tasting vegetable oil in the glass/foil container. Bring the water to a gentle simmer. The wax will melt very slowly. When melted, mix together the wax and oil using a pastry brush (preferably silicone — easier to clean). This mixture is commonly called “white oil”. (For a more detailed description, see the Wolfert article linked to above.)

Brush the insides of the very hot moulds thoroughly with the white oil, then invert them on a layer of paper towel placed on a cooling rack. Place a few layers of paper towel on an oven tray, transfer the metal moulds to the paper on the oven tray and put this in the oven for five minutes to let excess wax drain off. Remove the tray and tip the moulds a little to help final run-off. When the moulds are cool, place them in the freezer for at least an hour.

Silicone moulds: you can’t successfully wax these because of temperature and adhesion problems. Instead, butter the moulds (and only every second one — see why a few paragraphs down), making sure there’s a good layer of butter on the walls near the top of each mould (otherwise, the batter tends to stick during cooking, resulting in a ballooning of batter as it tries to rise, rather than a flat “shelf” of batter rising out of the mould). Place in the freezer for at least an hour.


Why freeze? It seems that the batter sticks a bit to the walls if the lubricant melts too quickly before the batter has formed an outer skin. As beeswax is solid until quite a high temperature, it helps prevent this sticking, even more so if in a frozen mould. Although silicone moulds are non-stick, they nonetheless aren’t adequately slippery without being buttered.


As mentioned earlier, it seems that a key step is to bake the tray of moulds over a strong heat source. So, for instance, if you have a lot of heat at the bottom of your oven (common for ovens with a concealed gas ring below), you place the tray on the bottom shelf. But if, like me, you have an old-fashioned oven with the gas flame at the back of the base, the heat rises up the back to make the top of the oven the hottest place. Not good. After much experimentation, I put a pizza stone in the centre of the oven, and heated my tray before placing the moulds on it and filling them. Result: perfection.

Moulds need to be filled with cold batter (stir well before pouring) up to about 2-3 mm below the rim of the mould. Metal moulds can be filled and then transferred to the hot tray (keep them quite spaced out), but silicone moulds are too unstable, so must be put on the tray and then quickly filled, leaving every second mould empty (silicone is a poor conductor of heat so you need good circulation of hot air).


Inadequate bottom heat seems to cause a sometimes massive rise out of the mould. The batter then bulges slightly and won’t sink back, yielding big bulbous burnt bottoms.

Cooking at 200C, conventional oven (ie, no fan), the batter should develop a slight fringe at 10-15 mins with the surface shimmering, by 20-30 mins it should be rising and rapidly protruding from the moulds, and at 45 mins it will have subsided and be about level with the rim of the mould. I like my canelés to be deep brown, with a strong odour of caramelisation wafting from the oven. It takes my oven about 1h15m to achieve this for the copper moulds and about 1h30m for silicone (by which time the top surface of the batter is charring badly — cover with foil earlier).

Your timing will depend on oven, moulds and personal preference. If you have a convection oven, I suggest dropping the temperature by 10-20C, but making sure the tray is close to the heat source (if at the bottom) or placed on a pizza stone.

The batter at the bottom of a silicone mould is very slow to brown, and the canelés as a whole tend to be more fragile, requiring a little resting time in the mould before removal. In contrast, canelés in copper moulds should be crisp straight from the mould, and fairly easy to get out (at worst, a decent whack on a benchtop should do it). Note that copper moulds stay hot for a very long time, and are very difficult to handle (I think rubber tongs or a slightly dampened tea towel would probably be best -if your towel is too damp you risk steam burns).


The ideal internal texture is soft, almost custardy, with a sort of open-crumb appearance like a loaf of bread, or sometimes described as a honeycomb pattern.

Canelés are best enjoyed still slightly warm, and certainly fresh (within a few hours of production). They go well with a dessert wine or strong coffee (as Wolfert also suggests), but they don’t really need any accompaniment. When no longer fresh, it’s best to refresh the canelés in a hot oven for a few minutes, then let them cool and crisp a little.

Finally, about those copper moulds… Although I’m all in favour of modern materials, the copper moulds really do produce a great product (and they’re soooo stylish;) ). I haven’t used aluminium ones. Silicone moulds are just trickier for a few reasons, but are an acceptable compromise if you see stars at the thought of coughing up the price of multiple copper ones (and I assure you, I’d love to have more than the paltry two I already own, but life goes on!).

Merry Christmas to all my readers!