What exactly happens when you're resting meat?

[Shalmanese]

It seems to me that resting is still resting firmly in the realms of culinary voodoo and nobody I've read has really given me a satisfactory explanation as to exactly what is happening inside the meat when it rests. Until today, I thought resting had to do with the fact that a piece of meat cooked over high heat was in thermal disequilibria, Because meat can hold on to differing amounts of liquid at different temperatures, the act of coming into equilibria would shift juices around the meat, some of which will leak if you cut it open.

But today, I cooked a rib roast using the Heston Blumethal inspired method of setting the oven at the desired temperature of the roast and letting the entire thing come into equilibria, under my logic, this would completely do away with the need for resting. However, when I actually cut into it, it was clear that it leaked far more juice that the well rested roasts I've cooked earlier.

So what actually happens during the rest that helps keep the juices in the meat and off the cutting board?

[jsolomon]

Whoo... that's a question that is a real poser. What happens when meat rests?

Well, first of all, you are asking questions about cooking under very different conditions. The first, where you cook at high heat and "thermal disequilibrium" is generally what chemists consider under kinetic control. Different pieces are under different controls. Even using something fairly simple like Le Chatelier's principle you can begin to fathom that there are simply a lot of chemicals doing a lot of things.

The second, Heston Blumenthal's method, is equilibrium control. Changes are more quantized as the whole of the chunk of material reaches temperatures at roughly the same time and then undergo roughly the same chemistry.

Equilibrium control is good for things like dissolving stuff. Kinetic control is good for producing many different compounds that are "metastable" compared to the equilibrium control products. Equilibrium control generally favors reversible reactions. Kinetic control generally favors irreversible reactions.

Beyond those generalities? That'd take some serious research.

[Patrick S]

The explanation that I have heard is that as meat cooks, muscle fibers in the meat contract, loosing their porosity, squeezing out their liquid, like a compressed sponge. The liquid doesn't leave the cut of meat necessarily, but it is pressed out of the fibers, into spaces outside of the fibers. So, if you cut the meat before the fibers have relaxed a little again, then all the liquid that is sitting in the extra-fiber spaces can drain right out, whereas if you wait a bit, and the fibers relax again, increasing the porosity of the fibers, then much of that liquid gets reabsorbed into the fibers, which holds it like a sponge. I don't know if that explanation is correct or not, but it is consistent with the observations you describe.

Edited July 14, 2005 by Patrick S (log)

[mizducky]

It seems to me that resting is still resting firmly in the realms of culinary voodoo and nobody I've read has really given me a satisfactory explanation as to exactly what is happening inside the meat when it rests. Until today, I thought resting had to do with the fact that a piece of meat cooked over high heat was in thermal disequilibria, Because meat can hold on to differing amounts of liquid at different temperatures, the act of coming into equilibria would shift juices around the meat, some of which will leak if you cut it open.

But today, I cooked a rib roast using the Heston Blumethal inspired method of setting the oven at the desired temperature of the roast and letting the entire thing come into equilibria, under my logic, this would completely do away with the need for resting. However, when I actually cut into it, it was clear that it leaked far more juice that the well rested roasts I've cooked earlier.

So what actually happens during the rest that helps keep the juices in the meat and off the cutting board?

←

Hmmm. I think you've got a few different mis-apprehensions going on here: (1) the problem that resting meat is supposed to address; (2) the way resting addresses it; and (3) the implications of "equilibrium" in a slow-cooking method such as Blumenthal's.

(1) First off, the problem which resting is meant to address, while it may well be exacerbated by disequilibria within the piece of meat, has a different root cause entirely. Patrick's got most of it--the process of applying heat to a chunk of meat causes the proteins within it to contract and be less capable of holding liquid. But in addition, the cooking process starts the breakdown/denaturing of those proteins--this in turn causes cell walls to leak or break down entirely, which releases even more liquid. Plus the overall heat being supplied to the meat raises its internal temperature, which is another way of saying the kinetic energy of the molecules within it increase. So as your roast approaches target-temperature (say, 140 F?) in your oven, you've got a chunk of meat with a lot of unfettered liquid, with proteins at decreased absorptive powers, and with a bunch of internal kinetic energy putting pressure on the internal liquid not unlike the pressure of water vapor within a teakettle on a hot burner, such that it's looking for any opening through which to push out.

As the meat continues to cook, the main things still holding the liquid in are: the ever-decreasing absorptive power of the proteins; the unbroken browned exterior of the chunk of meat; and the external kinetic pressure of the heated oven all around the meat, pushing back on the internal kinetic energy of the liquid in the roast. Yes, this is an equilibrium between that external and internal pressure. HOWEVER--once you remove the roast from the oven, the kinetic pressure from the oven's obviously gone away, and thus so has that equilibrium. Now it's mainly the unbroken crust doing the main holding action against the flood of internal liquid. So if you breach that unbroken crust right at the moment you remove the roast from the oven, yep, tons of juice will run out. Doesn't matter if you approached that target internal temperature fast or slow--what matters is the temperature differential between the interior of the roast and the ambient temperature of the room in which the roast is actually cut open.

2) So resting addresses the above problem by (a) allowing time for the roast's internal temperature to come down somewhat, decreaing the kinetic energy of all those interior juices; and (b) allowing time for the proteins in the meat to relax a little bit, so they can regain some absorptive power and do a better job of holding onto the juices. Result: you cut into the meat a half-hour after removal from the oven, and while there is some juice leakage, it isn't the wholesale gusher you would have gotten a half-hour previously. Which brings us to ...

3) "Equilibrium" in the context of slow-roasting a la Blumenthal or other slow-cooking advocates: while I'd think these are definitely excellent methods for getting uniformly-cooked and tender roasts, I don't think the equilibrium they induce reduces the need for resting. First-off--what is meant by "equilibrium" in this context anyway? Because there's a whole bunch of different equilibria going on, in terms of heat, energy, pressure, chemical processes, etc. I think the most important ones in this context have to do with temperature/energy/kinetic "pressure," but even then, you've got multiple equilibria going on--the chunk of meat has an internal equilibrium, but there's also an equilibrium between the meat and its environment (be that oven, room, refrigerator, or whatever). So it's important when talking about equilibrium here to distinguish which one you're talking about.

Secondly, as I noted above, the problem that makes resting necessary has nothing to do with dis-equilibria *within* the meat, but with the results of applying heat to meat in general, plus the dis-equilibrium between the meat's interior and its environment caused by removing said meat from the oven. And a perfectly-cooked target-temperature roast, whether Blumenthal-method or otherwise, is only in temperature-equilibrium with its environment as long as it's still in the oven. Once you remove the roast from the oven to that presumably 70-deg (or whatever ) room, that equilibrium's gone, replaced by a temp/kinetic-energy differential between roast interior and ambient temperature. If you cut into that roast right at that moment, Blumenthal method or no, it will still bleed out its juicy goodness. So yep--you still have to rest the puppy, and it has nothing to do with how perfectly you followed Blumenthal's method.

Now mind you, I can imagine that if you roasted a chunk of meat, either quickly or slowly, long enough for significant amounts of liquid to evaporate out of the meat, there would be no significant liquid leakage even if you cut into it seconds after removal from the oven. But I don't think I'd really want to eat that piece of shoe-leather.

[okay ... now I post, and wait for someone to find the inevitable hole in my logic ... ]

[russ parsons]

a most convincing answer msduck. correct me if i'm summing it up incorrectly:

1) it allows the free juices within the meat to redistribute more evenly throughout.

2) it allows reserved heat in the meat to continue cooking to the desired doneness (sort of the equivalent of coasting to a stop, because you certainly can't slam on the breaks with a roast).

[mizducky]

a most convincing answer msduck. correct me if i'm summing it up incorrectly:

1) it allows the free juices within the meat to redistribute more evenly throughout.

2) it allows reserved heat in the meat to continue cooking to the desired doneness (sort of the equivalent of coasting to a stop, because you certainly can't slam on the breaks with a roast).

←

Actually, I don't think I really addressed your point #2 as such--when the proteins relax again, that's not so much from further cooking but from removal of heat energy. But yeah, there is enough residual heat energy that the meat does continue cooking for awhile even out of the oven until that excess heat radiates off. (That's the same reason why it's best to remove an omlette from the pan while it's still a tad runny, as residual heat will continue cooking it; otherwise, if you waited until it was perfectly done to removed it from the pan it would wind up being rubbery-overdone once it sat on the plate for even a couple of minutes.)

There's also a point I neglected to address, about whether that browned surface on the roast really does hold the juices in. I've read the authorities who debunk the concept that searing meat surfaces holds juices in. It just doesn't match my observations that breaching that surface starts the leakage. Maybe I need a little elucidation on that point myself. In the meantime, though, I'm gonna stick with what I've observed.

[Shalmanese]

(1) First off, the problem which resting is meant to address, while it may well be exacerbated by disequilibria within the piece of meat, has a different root cause entirely. Patrick's got most of it--the process of applying heat to a chunk of meat causes the proteins within it to contract and be less capable of holding liquid. But in addition, the cooking process starts the breakdown/denaturing of those proteins--this in turn causes cell walls to leak or break down entirely, which releases even more liquid. Plus the overall heat being supplied to the meat raises its internal temperature, which is another way of saying the kinetic energy of the molecules within it increase. So as your roast approaches target-temperature (say, 140 F?) in your oven, you've got a chunk of meat with a lot of unfettered liquid, with proteins at decreased absorptive powers, and with a bunch of internal kinetic energy putting pressure on the internal liquid not unlike the pressure of water vapor within a teakettle on a hot burner, such that it's looking for any opening through which to push out.

As the meat continues to cook, the main things still holding the liquid in are: the ever-decreasing absorptive power of the proteins; the unbroken browned exterior of the chunk of meat; and the external kinetic pressure of the heated oven all around the meat, pushing back on the internal kinetic energy of the liquid in the roast. Yes, this is an equilibrium between that external and internal pressure. HOWEVER--once you remove the roast from the oven, the kinetic pressure from the oven's obviously gone away, and thus so has that equilibrium. Now it's mainly the unbroken crust doing the main holding action against the flood of internal liquid. So if you breach that unbroken crust right at the moment you remove the roast from the oven, yep, tons of juice will run out. Doesn't matter if you approached that target internal temperature fast or slow--what matters is the temperature differential between the interior of the roast and the ambient temperature of the room in which the roast is actually cut open.

Is this the whole thing about the sealed exterior "sealing in the juices"? I think it's been quite conclusively debunked that a seared exterior does precisely nothing at all in holding in interior liquids. The crust is porous enough such that it's no barrier at all. And what do you mean you need the roast to come into thermal equilibria with the room? You want the meat to come completely to room temperature? That doesn't sound very appetizing, besides, most people routinely serve well-rested meat thats far above room temperature.

I'm not convinced by this explaination

2) So resting addresses the above problem by (a) allowing time for the roast's internal temperature to come down somewhat, decreaing the kinetic energy of all those interior juices; and (b) allowing time for the proteins in the meat to relax a little bit, so they can regain some absorptive power and do a better job of holding onto the juices. Result: you cut into the meat a half-hour after removal from the oven, and while there is some juice leakage, it isn't the wholesale gusher you would have gotten a half-hour previously. Which brings us to ...

This may very well be correct and would at least be consistant with observations. It may be that meat at 55C will hold much more liquid than meat at 60C, in that case, letting it rest for the requisite n minutes will decrease the internal temperature enough such that all the juices are absorbed into the fibre. If this is the case, then a far superior way of resting would be to first cook the meat in a 60C oven until the internal temperature hits 60C, then to drop the oven down to 55C and wait again until the internal temp hits 55C. In this case, the entire chunk of meat becomes equally rested which is far superior to the benchtop method of resting where the exterior becomes over rested (and thus, too cold) and the interior becomes under-rested. However, experimentation would be needed to determine exactly how much of a temperature drop produces well rested meat.

3) "Equilibrium" in the context of slow-roasting a la Blumenthal or other slow-cooking advocates: while I'd think these are definitely excellent methods for getting uniformly-cooked and tender roasts, I don't think the equilibrium they induce reduces the need for resting. First-off--what is meant by "equilibrium" in this context anyway? Because there's a whole bunch of different equilibria going on, in terms of heat, energy, pressure, chemical processes, etc. I think the most important ones in this context have to do with temperature/energy/kinetic "pressure," but even then, you've got multiple equilibria going on--the chunk of meat has an internal equilibrium, but there's also an equilibrium between the meat and its environment (be that oven, room, refrigerator, or whatever). So it's important when talking about equilibrium here to distinguish which one you're talking about.

Secondly, as I noted above, the problem that makes resting necessary has nothing to do with dis-equilibria *within* the meat, but with the results of applying heat to meat in general, plus the dis-equilibrium between the meat's interior and its environment caused by removing said meat from the oven. And a perfectly-cooked target-temperature roast, whether Blumenthal-method or otherwise, is only in temperature-equilibrium with its environment as long as it's still in the oven. Once you remove the roast from the oven to that presumably 70-deg (or whatever ) room, that equilibrium's gone, replaced by a temp/kinetic-energy differential between roast interior and ambient temperature. If you cut into that roast right at that moment, Blumenthal method or no, it will still bleed out its juicy goodness. So yep--you still have to rest the puppy, and it has nothing to do with how perfectly you followed Blumenthal's method.

←

This seems to be largely the thrust of the first point and I still remain unconvinced. My main problem with it is to view the resting process from the perspective of the centre of the meat. How does the roast know where it is after 10 minutes or 50 minutes? A chunk of meat doesn't automatically know what the temperature is 30cm away from it, there needs to be some mechanism to convey to the centre of the roast that it is now sitting on the table and not in the oven. Under your theory, it seems that resting should increase fluid loss, not decrease it. After all, immediately out of the oven, it has a completely flat temperature gradient from the inside to the outside of the roast. 30 minutes out of the oven, it has a very steep temperature gradient as the core is still at 60C while the outside will be at 30C. This would imply that there will be a larger migration of fluids within the body after 30 minutes of resting and, thus, should produce more fluids when cut.

Personally, I think by the process of elimination, cause B is the most likely to be what the correct answer is. The butcher is selling aged rib roasts criminally cheap so I will buy another roast next weekend and try the method 2 of roasting to 60C and then reducing the oven to 55C until the internal temp hits 55C. I will report back with my findings.

Edited July 14, 2005 by Shalmanese (log)

[Shalmanese]

2) it allows reserved heat in the meat to continue cooking to the desired doneness (sort of the equivalent of coasting to a stop, because you certainly can't slam on the breaks with a roast).

←

But this ignores the fact that I wasn't using the conventional high heat roasting method. Under the Blumethal method, there is no coast so this should not be a factor.

[russ parsons]

After all, immediately out of the oven, it has a completely flat temperature gradient from the inside to the outside of the roast. 30 minutes out of the oven, it has a very steep temperature gradient as the core is still at 60C while the outside will be at 30C.

←

this seems to be the opposite of my understanding. The surface of the roast is always much, much hotter than the center (depending on the size of hte roast and hte heat of the oven, of course).

on the other hand, it seems to me that after being pulled from the oven, the surface and outer perimeter would cool at a much quicker rate than the center, resulting in more even doneness.

this theory has the added advantage of fitting with practical experience: But think of a slice of prime rib: the surface will be browned (300-degrees plus), the outer perimeter will be medium to well (160 plus) while the center is still pink (130). Granted, cooking a smaller roast in a cooler oven will result in more even doneness, but there will still be a range.

As for the "sealing in the juices", i'm not going to touch that can of worms, except to point out that, practically speaking again, if you slice the meat right out of the oven, you end up with a lot more juice on the board than if you let the meat rest.

[Shalmanese]

Equilibrium control is good for things like dissolving stuff.  Kinetic control is good for producing many different compounds that are "metastable" compared to the equilibrium control products.  Equilibrium control generally favors reversible reactions.  Kinetic control generally favors irreversible reactions. 

This is an interesting point but are there really any desirable meta-stable compounds between 60C (medium rare) and say, around 130C (when maillard reactions start in earnest). In my blowtorch thread, I pointed out that I got around the need for high heat searing and browning by using a blowtorch which would only bring the very skin of the meat above 60C and leave the rest of the meat completely untouched. Thus, I have a very thin region that produces the meta-stable maillard chemicals that are needed for browning and I have the rest of the meat at 60C and in thermal equilibria, is there anything I'm missing out on in terms of chemicals?

I know collagen starts dissolving at just below 60C but it takes quite a while for it to proceed at that low tempeature so it was quite likely that my roast did not have as much gelatine as it could have. Blumethals method requires 20 hours in the oven, presumably to fully dissolve the collagen but a rib roast seems to not have much connective tissue at all so I was willing to live with that.

In short, is there any disadvantage to the blowtorch/low oven way of cooking compared to the conventional way apart from the time it takes?

[Shalmanese]

After all, immediately out of the oven, it has a completely flat temperature gradient from the inside to the outside of the roast. 30 minutes out of the oven, it has a very steep temperature gradient as the core is still at 60C while the outside will be at 30C.

←

this seems to be the opposite of my understanding. The surface of the roast is always much, much hotter than the center (depending on the size of hte roast and hte heat of the oven, of course).

on the other hand, it seems to me that after being pulled from the oven, the surface and outer perimeter would cool at a much quicker rate than the center, resulting in more even doneness.

this theory has the added advantage of fitting with practical experience: But think of a slice of prime rib: the surface will be browned (300-degrees plus), the outer perimeter will be medium to well (160 plus) while the center is still pink (130). Granted, cooking a smaller roast in a cooler oven will result in more even doneness, but there will still be a range.

As for the "sealing in the juices", i'm not going to touch that can of worms, except to point out that, practically speaking again, if you slice the meat right out of the oven, you end up with a lot more juice on the board than if you let the meat rest.

←

That would be true in conventional roasting. However, the Blumethal method (which I'm just going to call LTLT for Low Temperature, Long Time from now on) is completely different.

Under LTLT, you set the temperature of your cooking apparatus to the desired temp of your meat. So when I took the meat out of the oven, the outside was exactly 60C and the centre was also exactly 60C. However, even under this method, there was significant leakage from the meat when it was taken straight out of the oven, suggesting that the "thermal gradient" idea of resting, while possibly a very strong factor in leakage, is not the only factor. Clearly, something else is going on here as well.

[RETREVR]

Talk about making things more complicated than they need be.

Think of a bell curve. The peak of which is 140deg. You pull the roast before it gets there which allows the heat of the outer portion of the roast to bring the temperature of the interior to the top of the curve. The muscles tense-up with exposure to heat(remember the trick of feeling your palm when relaxed for rare and feeling it while flexed for well). Same thing. You have to wait untill the temp starts sliding down the other side of the curve for the muscles to relax again....thus your roast won't wet the bed.

I can almost see why you thought that you could avoid this with your coast method. However, the roast was riding the top of the bell curve and had not yet started to cool. I can tell you that I can put a 145deg roast in a 145deg holding cabinet and it will leak its juices even without being cut. To avoid this I use a 120deg cabinet to hold a roast(not for an extended amount of time but 30-40 minutes)

I use a small cut of beef for coctail sandwiches all the time. I used to roast in the morning and wait untill the afternoon to slice it. Now I roast the day before cool them overnight before slicing. I get almost no loss of juices.

[Shalmanese]

Whats the difference between a 145F holding cabinet and a 145F oven? One key feature about the LTLT method is that no juice leaks from the roast, not even a single drop. Which means an extra juicy roast but no pan drippings. So I don't see why a holding cabinet would behave differently from the oven. Maybe the act of handling the roast would cause it to leak if your squeezing it or putting pressure on it. And it sounds like your 120F holding cabinet idea is the same as my proposal. The process of "resting" happens when a roast drops a fraction from it's peak temperature.

[Shalmanese]

Success, I bought another standing rib roast and tried my proposed method.

I set the oven to 60C and waited until the core reached 60C which took about 5 hours. Then, I set the oven to 55C and waited until the core dropped to 55C. This took another 3 hours and it certainly isn't for the impatient cook. Luckily, I started the entire thing early in the morning and had an entire day to cook it.

And the best part, not a single drop of juice came out of the roast, no matter how I mauled or squeezed it. I'm now convinced that at least one of the factors behind resting is that the meat must be allowed to drop at least X degrees below the peak temperature which allows the muscles to relax enough to absorb free juices. This means instead of relying on mystical voodoo about waiting 30 minutes per pound or whatever nonsense, you can just rely on a temperature probe and carve when it hits the desired temperature. You can even rest on the bench which will give a faster rest but cooler outsides.

[Ann_T]

Seems complicated.

I always thought that the simple explanation was that as the meat roasted and the juices got hot they bubbled to the surface and that the purpose of letting the meat rest was so that the juices had a chance to slowly be absorbed back into the meat. We like our beef on the rare side I always take it out of the oven around 120 so that while it rests the temperature rises to between 125 and 130°F.

Ann

[slkinsey]

"Resting" seems to be one of those things that is necessary for conventional cooking methods, where the cooking environment is substantially hotter than the target temperature of the meat. But even in the LTLT method Shalmanese propounds, it would seem that resting is somewhat necessary, so there has to be something to the idea that a reduction from the meat's peak temperature is beneficial with respect to cooking method regardless of cooking method. Whether something like Shalmanese's method LTLT cooking followed by another LTLT reduction in temperature constitutes "resting" is perhaps another question entirely.

[orangeblossom]

Is there a difference between a roast which is foil-tented and left to rest, and leaving the roast in the hot oven (switched off) with door ajar?

Thanks.

Edited November 30, 2005 by orangeblossom (log)

[andiesenji]

Is there a difference between a roast which is foil-tented and left to rest, and leaving the roast in the hot oven (switched off) with door ajar?

Thanks.

←

You should bring it out of the oven, uncover it if tented, and let it "coast" at room temperature.

The difference in heat between the inside of the roast and the room temp, is what allows the meat to reabsorb the juices that have been pushed to the outside during the roasting.

Note these two ways of roasting. Only the high temp one does not require coasting outside the oven.

[jackal10]

If you roast using a long time low temperature method (e.g. 8-12 hours at 140F) you don't need rest at all since the juice is never squeezed out

[ned]

In the oven with the door open is still a hot place. Your roast will continue to cook some. How much? Hard to know. It'll also continue to cook a little on the counter because of the residual heat in the roast itself.

Ultimately I'd say it's better to rest on the counter. That way you have more control.

[carswell]

It all depends. My kitchen is cool in the winter and drafty in the summer, so leaving the roast on the counter often results in its cooling a bit quicker than is ideal. My solution is to leave the roast on the counter for a few minutes while the oven, with its door ajar, cools off, then to place the roast back in the oven. A big advantage of this approach is that it frees up counter space, which I don't have a lot of.

[ChefCrash]

It all depends. My kitchen is cool in the winter and drafty in the summer, so leaving the roast on the counter often results in its cooling a bit quicker than is ideal. My solution is to leave the roast on the counter for a few minutes while the oven, with its door ajar, cools off, then to place the roast back in the oven. A big advantage of this approach is that it frees up counter space, which I don't have a lot of.

←

This sounds reasonable.

[monavano]

If you use the high temp method, i'd recommend bringing it out onto the counter for 15 minutes give or take. normal carry-over cooking, out of the oven, will cause the temp to rise 5-7 or so degrees.

last Christmas, we went to my in laws for a brunch. i was put in charge of making a whole filet mignon (which i cut in half). After pan browning, i placed in in a very hot oven and I used my MIL's digital thermometer and set it to beep at 130 degrees. Well, I didn't set it right,or perhaps the beeper didnt work. I fortunatly caught the smaller piece at 135 and immediately pulled it out as well as the bigger one.

Knowing I wanted med rare to med, I placed them on a serving plate and wisked the filets into the garage and placed them on the freezing cold floor for 15 minutes.

Turned out beautifully.

Edited November 30, 2005 by monavano (log)

[orangeblossom]

It all depends. My kitchen is cool in the winter and drafty in the summer, so leaving the roast on the counter often results in its cooling a bit quicker than is ideal. My solution is to leave the roast on the counter for a few minutes while the oven, with its door ajar, cools off, then to place the roast back in the oven. A big advantage of this approach is that it frees up counter space, which I don't have a lot of.

←

Yes, counter space is a concern for me too, esp when I'm cooking for friends, which was why I raised the question

[Shalmanese]

If you roast using a long time low temperature method (e.g. 8-12 hours at 140F) you don't need rest at all since the juice is never squeezed out

←

Althought this sounds intuitive, experimentation proves that it is absolutely untrue. LTLT cooked meat leaks almost as much as conventional, high temp roasting with inadequate resting. However, what appears to stop leakage is to allow the meat to cool down a few degrees from peak. Under a traditional LTLT approach, this would involve cooking the meat for 12 hours on 140F until the centre hits 140F and then resting for another 6 hours on 135F until the meat cools down to 135F. Alternatively, using a high temperature method, you could try cooking the roast at 350F until the centre hits 130F, taking it out to rest on the counter top, letting it coast up to 140F and then dropping down to 135F. This will lead to less even cooking and a slightly cooler roast on average but takes less time.