Dr. Weiss and Dr. St-Pierre co-authored this episode’s journal club paper in Applied Animal Science (ARPAS Journal). Bill and Normand share a career-long interest in how feedstuffs and diet variation impact cows. (6:31)
Dr. Weiss and Dr. St-Pierre co-authored this episode’s journal club paper in Applied Animal Science (ARPAS Journal). Bill and Normand share a career-long interest in how feedstuffs and diet variation impact cows. (6:31)
Bill and Normand discuss sources of variation, which they divide into true variation and observer variation. True variation means the feed has changed: a different field, change during storage, etc. Observer variation includes sampling variation and analytical variation. Some feeds may exhibit a lot of true variation and others may exhibit a lot of observer variation. And some feeds are high in both types of variation. Highly variable feeds should be sampled more frequently. Some feeds are so consistent that using book values makes more sense than sending in samples for analysis. Bill and Normand go on to give some examples and share sampling and analysis tips for different types of feedstuffs. (12:41)
Bill would often be asked if users should continue to average new samples with older ones or just use the new numbers from the most recent sample. He and Normand debate the pros and cons of the two approaches as well as discuss the use of a weighted average where recent samples would be weighted to contribute more. (26:02)
Next, our guests discuss how multiple sources of a nutrient reduce the TMR variation for that specific nutrient. For example, alfalfa NDF is more variable than corn silage NDF on average. Yet if you use a blend of these two ingredients, you end up with less variation in NDF than if you used all corn silage. Normand details the mathematical concepts behind this relationship. Both Bill and Normand emphasize that diets must be made correctly for the best results. (32:26)
How do feedstuffs and diet variations impact cows? Both guests describe different experiments with variable protein and NDF concentrations in diets. Some were structured, like alternating 11% CP one day and 19% CP the next for three weeks. Some were random, like randomly alternating the NDF over a range of 20-29% with much higher variation than we’d ever see on-farm. The common thread for all these experiments is that the diet variations had almost no impact on the milk production of the cows. (38:04)
Clay asks how variation in dry matter might affect cows. Bill describes an experiment where the dry matter of silage was decreased by 10 units by adding water. Cows were fed the wet silage for three days, twice during a three-week study. To ensure feed was never limited, more as-fed feed was added when the wet silage was fed. It took a day for cows on the wet silage treatment to have the same dry matter intake (DMI) as the control cows and milk production dropped when DMI was lower. However, when switching abruptly back to the dry silage diet, DMI increased the day following the wet silage and stayed high for two days, so the cows made up for the lost milk production. Bill and Normand underline that it is critical for the cows not to run out of feed and described experiments where feed was more limiting, yielding less desirable outcomes. (46:17)
In the last part of the paper, our guests outlined seven research questions that they feel need to be answered. Normand shares that his number one question is how long will cows take to respond to a change in the major nutrients? He feels that we spend an inordinate amount of money on feedstuffs analysis, and there are some feeds we should analyze more and some feeds we should quit analyzing. Bill’s primary research question revolves around controlled variation. What happens if you change the ratio of corn silage and alfalfa once a week? Will that stimulate intake? Data from humans, pets, and zoo animals indicate that diet variation has a positive impact and Bill finds this area of research intriguing. (50:43)
In closing, Clay encourages listeners to read this paper (link below) and emphasizes the take-home messages regarding sampling and research questions. Normand advises that if you are sampling feed, take a minimum of two samples, and try as much as you can to separate observer variation from true variation. He also reminds listeners to concentrate on a few critical nutrients with more repeatability for analyses. Bill encourages nutritionists to sit down and think when they get new data - before they go to their computer to make a diet change. If something changed, why did it change, and is it real? Take time to think it through. (1:01:38)
You can find this episode’s journal club paper from Applied Animal Science here: https://www.appliedanimalscience.org/article/S2590-2865(24)00093-4/fulltext
Please subscribe and share with your industry friends to invite more people to join us at the Real Science Exchange virtual pub table.
If you want one of our Real Science Exchange t-shirts, screenshot your rating, review, or subscription, and email a picture to anh.marketing@balchem.com. Include your size and mailing address, and we’ll mail you a shirt.
Scott Sorrell (00:00:09):
Good evening everyone, and welcome to the Real Science Exchange, the pubcast where leading scientists and industry professionals meet over a few drinks to discuss the latest ideas and trends in animal nutrition. Hi, I'm Scott Sorrell, gonna be your host here tonight at The Real Science Exchange, and tonight we've got two great guests. We're back with our next episode of the Journal Club, and as always, we got Dr. Bill Weiss here, emeritus professor at the Ohio State University. Now, Bill, when I first saw that we had this this, this podcast scheduled to record right after the game this Saturday and that I was gonna be sharing some time with a couple buck guys, I thought it would be fun to kind of reminisce about the victory. It's not gonna happen, is it?
Dr. Bill Weiss (00:00:53):
It would be, it would be fun to reminisce, but, but unfortunately, maybe next year.
Scott Sorrell (00:00:59):
So yeah, maybe next year. That's what we've been saying for about four years. But you know, at least I'm still a Penn State fan. They're still, they're still well and alive, so looking forward to watching some Penn State football. Anyway, Bill, I'm really looking forward to this podcast, very practical in nature. And why don't you give us just a real top line overview of what that is and the, and, and then introduce the guest that you brought with you tonight.
Dr. Bill Weiss (00:01:26):
This is a little different this time because I'm one of the authors the paper was published in Applied Animal Science, which is the A journal, and I'll put a plug in for Harus. If you're not a member, you should really think about joining. It just came out in the, the most recent edition, I think that's November, but it might be October. It's the most recent one. And the paper is called Perspectives and Commentary variation in Nutrient Composition of Feeds and Diets, and how it can affect Formulation of Dairy Cow diets. My guest here is a longtime colleague and friend. We worked on this project for probably 10 years, and that's Dr. Norman St-Pierre, previously from Ohio State University. And now happily a sailor. So welcome, Norman.
Dr. Normand St-Pierre (00:02:15):
You're welcome, Bill.
Scott Sorrell (00:02:16):
Yeah. Good to have you here Nor. Now, Bill mentioned that you were a sailor and, and I'm fascinated Norm, by where you live, and I think our audience would find it equally as fascinating. Why don't you kinda give us an idea of, of where you live, kind of what that environment's like and, and what you're spending your time doing these days.
Dr. Normand St-Pierre (00:02:34):
Oh, okay. Well, I'm retired now, fully retired, and I lived in Virginia, but it's a small island called Tengier, T-E-N-G-I-E-R. Tengier Island was named Sway by Captain Smith when he was exploring Chesapeake Bay, because we have a long beach that resemble a bit like the beach we have in Morocco, in in Tangia Morocco. Well, anyway, so it's that little island. There's that 12 miles of water, pretty much surrounding all around about 400 residents remaining. And we have kindergarten to 12 grade school, about 50 students in total. Very calm. If the world comes to an end, come over here with 50 years behind. So you have, you have 50 years to keep going on. The, the, it's very peaceful, very slow pace. And so I have a sailboat. I love sailing. And so that's where the boat is. It's about two minutes away from where I live. And so any morning when I get up and there's a nice breeze, it doesn't matter what it rains or not, it doesn't matter if it's cold or not, there's a nice breeze. I'll be out on the water within and can to plan for the day. I think everything can wait for tomorrow nowadays. So that's where I'm at.
Scott Sorrell (00:03:47):
Good. And that's right in the middle of the Chesapeake Bay or close?
Dr. Normand St-Pierre (00:03:50):
Yeah, yeah, right in the middle, pretty much South Chesapeake Bay. If you follow the Potomac River just keep going east and you're gonna run just about in our little island, there's no bridge, of course. There's some, some kind of ferry, but not for cars. Most people on the island drive a golf cart. Either that or an old beaten up pickup trucks.
Scott Sorrell (00:04:13):
Yeah. Wow.
Dr. Normand St-Pierre (00:04:14):
That's where it's at.
Scott Sorrell (00:04:15):
So, yes. Sounds like fun. So I want to thank Bill for inviting you to join us tonight. Hey, Bill, gotta tell me what's in your glass tonight?
Dr. Bill Weiss (00:04:22):
Well, it's a, a, a beer and it is a special occasion beer. I'm not making this name up. It's from 50 West Brewery and it is called Wiener Mobile. And so what I saw that name, I had to buy it, so, okay.
Scott Sorrell (00:04:39):
And what was the special occasion?
Dr. Bill Weiss (00:04:42):
It, well, with Norman, Wiener Mobile, he likes to sing the Oscar Meyer song, so that's Wiener Mobile for that. So
Scott Sorrell (00:04:49):
Well, I also like to introduce my co-host for this evening, Dr. Clay Zimmerman. Clay, welcome back to the Real Science Exchange. It's been a while.
Dr. Clay Zimmerman (00:04:58):
Yep. Thanks Scott.
Scott Sorrell (00:04:59):
And do you have anything special tonight in your glass?
Dr. Clay Zimmerman (00:05:02):
I have some, a hard cider here.
Scott Sorrell (00:05:04):
Hard cider again. Alright. I do very well. How about you?
Dr. Clay Zimmerman (00:05:07):
What do you drink?
Scott Sorrell (00:05:08):
Well, I'm just going with my, my forever bourbon. I, i, I need to get out of a rut and, and go find some good stuff. But that, that's all I've got. It was, it was a long weekend. So anyway, Hey guys, I'm really looking forward to this one. So here's to a great podcast. Cheers.
Commercial (00:05:35):
The economics of Feeding ReaShure Precision Release Choline. ReaShure fed during the transition period. And because it's fed for such a short period of time, it costs just $15 per cow, and yet the benefits will continue to generate income. Throughout the year, cows fed er produced five pounds more colostrum, which pays for your reassure investment on the very first day of lactation cows Fed reassure also produce five pounds more milk per day every day. That means after the first day, every day is payday, invest in ReaShure during the transition period and recoup your investment on the very first day of lactation after that. You got it? Payday. All right, Bill, why don't you just jump right into it and kinda lead us into the discussion.
Dr. Bill Weiss (00:06:31):
Okay. Well, first thi this paper is different than a lot of 'em because it's perspective and commentary, so it's, it's kind of a review, but the authors are given more freedom to speculate a bit and try and interpret the data a little more broadly than you would for a a, a regular scientific paper. And yes, I'd like to start off with Norman and just, you know, this is something we've worked on for, I think we published the first paper in 2012, but why, why do you, why do we do this? Or why do we start this project? Do you remember that far back?
Scott Sorrell (00:07:05):
Great question.
Dr. Normand St-Pierre (00:07:05):
Well, yes, I, I do some, because basically you have to go back to my PhD. That, that's when I started looking at nutrition and nutrition of cows with the perspective things vary. So how should we control that? And at the time, it was through diet formulation, stochastic programming was basically what came outta that. But all throughout this, there were questions that suddenly you start asking, so people sample forages, how often? Well, basically the answer, well, you go and you see no nutritionist visit the herd once a month. You sample the forage, you get the results, you plug those in your computer, you issue new diets. And I started wondering, so why, if nothing changed, why would you do that? And if it changes in between the 30 days and the month, what happens? Unclear what happens to cows when you have some, some variation.
Dr. Normand St-Pierre (00:08:00):
And so what we started doing, I'd say about 15 years ago, Bill, was to actually run some trial and looking at what does variation what does it do to cows? Before that, I was looking at from an eco economic basis, if, if you sample, it becomes a quality control issue. And so as you sample, things might not have changed, but unless you sample and analyze, you won't find out when it changes, and therefore the price that you'll be paying will be the loss of productivity. On the other hand, if you never want to do anything wrong, you would test and analyze every pound going in the mixer wagon would cost you a fortune, but my all direction would be always perfected. And so at some point you're gonna have a happy compromise that. And so up the model to the in there and basically ended up a concluding in general that first people don't do it right because they take only one sample and you have no idea about variation.
Dr. Normand St-Pierre (00:09:03):
Then remember in, in my courses in statistics, you learn that you never gain more information on the variance of something until you go from an end of one to two at end of one. You have no idea. It can be infinity. You have no idea whatsoever. When you go in NF two, you don't have a good estimate, but at least you do have an estimate of the variation. And so people, when they go and sample something should at least take two samples minimum, because at that point, that's the only way you can at least start looking at. But two independent samples, and I have seen people going and, and going in a bunker silo and you unload a a bunch and then take a handful and pail fill up a pail, and then they take a sample and they put it in one bag and taking on a sample, put another bag.
Dr. Normand St-Pierre (00:09:52):
All that was gonna tell them is, is how well they were how good they were at sub sampling the bucket that they had. And so that's, these two samples are not independent. The best way would be a nutritionist walks on a farm today, sample the silage. The first thing is he arrives there, then does his routine stuff, and then come back and take another sample. So that's the very first thing that, that we know that if you're gonna do something that's, that's a, that's a rule. You're gonna do something, do it twice independently, because that's the only way you gain information on the areas. The work call. So that we did clearly started looking at, there's many factors that cause variation, apparent variation in the results that you get. And some of those are intrinsic to the feed. The feed has changed and therefore the analytics should change.
Dr. Normand St-Pierre (00:10:44):
But also any sampling and sub sampling in any analytical assay will have some inherent variation. We call that error. An error is not quite the right term. It should really be pre precision or imp precision or repeatability. And so you have assays in the lab that are quite repeatable. A KOL assay, for example, is quite repeatable, will have a ion of variation of one, one and a half. Biological assays, if you're trying to incubate doing in C two those will have a coercion of variation of at least 5%. And, and so at that point, going, getting one result out of one sample and even two samples, and you're looking at that number and lab labs contribute to the fraudulent interpretation there say fraudulent because they will report 85.67%. Postal digestibility, 85.67 only means it's anywhere between 80-90%. That's what so start reformulating because the number has changed by half a percent.
Dr. Normand St-Pierre (00:11:57):
This is hogwash. You're wasting time, money and in, in a ways, you know, when in when you have a hammer, everything looks like a nail. The nutritionist looks at things all like in nutrition and the, the, I don't mean that they're doing that purposefully, but they feel like they have to change the diet because the number has changed and the feed may not have changed at all. So the initial question, so when and why? It's because basically it had been very, very little work done on the variation, what causes the variation and how it does affect cows. And to this day, there's still a lot of questions are unanswered with them.
Dr. Bill Weiss (00:12:41):
In, in this paper, we, we spend a lot of time talk, talking about sources of variation. And now in English, not statistic E what, why do users need to know the, you have a standard deviation of five or whatever, why do they need to know what contributed to that?
Dr. Normand St-Pierre (00:13:02):
Well, because of all the sources, all the factors that cause variation, you can group them in, in two groups. So you have, in total, you have your apparent variation, how apparently the results are, are changing and it's gonna be made of true variation. That is the feed has changed. It can be from a different field, it can be from whatever factors have had contributed to the change, including change during the the storage. The start digestibility of corn silage will change through time. So these are real, and, and if you do believe that cows need to have balance diets, then you need to balance for those. And the real variation will be changing the expected value even with what the ration will be delivering. So these are real, but on the other hand, you have all the other ones that are the factors that are cause apparent observer variation.
Dr. Normand St-Pierre (00:14:03):
And, and what it means here is it's the feed hasn't changed at all. It just, the results due to variation, it's easy to understand. For example, sampling variation. There's some feed that are really hard to sample. Corn silage is one of those. 'cause They're made of particles of different size in each size particles and very different composition. So, and you're trying to put that in 200 grand in a little bag, you're gonna have some, some sampling variation in there. Cotton seed on the other hand, has almost no sampling variation. You grab one seed, you have 150 seed coming with it, put that in the bag and, and you can repeat. The process won't change very much. You have analytical variation that anybody who's working a lab will understand that. Now in research, we take samples and we always run it at least twice, generally three times.
Dr. Normand St-Pierre (00:14:52):
And then we'll look at the results. And that's how we can say out of the same samples, and we have hundred 50 samples out of that block. The variation that we have, the analytical variation is X percent. So for example, earlier I said KOL, we see typically one, one and half, maybe 2% as a standard devi deviation. So that means if I'm saying, if I were to see repeatedly the same sample of alfalfa hay to a lab, and the true value of crude protein is 20%, the results coming from the lab will vary easily between 19 and 21%. And so if last month the result was 19 and a half and this month is 20, 20 and a half, doesn't mean it has changed. Just very apparent variation that you have in the lab. If compare that at some point in time, I remember I was in Australia and they decided to take me out on a kangaroo hunt.
Dr. Normand St-Pierre (00:15:54):
Well, you, you stand in the back of a pickup truck and you go on on the road, which is is all bumping everything and the kangaroo is jumping. You're trying to shoot that thing. Well, if you have a scope and you look in there, the kangaroo is moving a whole lot more than what the kangaroo is actually moving because you're moving too. And so you moving is the, the observer variation and the kangaroo is the true variation that you have. And of course I didn't, I wasn't very successful, so I hope nobody was hoping to eat kangaroo steak that night. And so it's very important because the proportion of those two or two group of factors vary a lot across different feeds. So for example, I've already alluded to whole cotton seed, you have it. The, the true seeds themselves have all very, very little variation.
Dr. Normand St-Pierre (00:16:47):
True variation in whole cotton seed almost does not exist. It's very small in fact. But if you go and take sample and send that to the lab, what is it they have to do? Well, the first thing, you have to grind it. Now you have three different substrate that you're grinding. You have the fuzz, you have the hulls, and you have the almond, the, the true seeds inside, and you grind that up. And then tell me how you're gonna be sampling that to put in that little machine. You need 0.2 grams of that to put to put in there. So that's where most of the errors will be. And so when we look at whole cotton seed and the result from the lab, it would show that the NDF no average is 35%, but it run anywhere between 30 and 40. Most of that is apparent variation in the lab.
Dr. Normand St-Pierre (00:17:34):
The true cotton seed probably didn't vary by 1%. Good. And so that's, that's a case where a a feed is actually quite constant in composition. You observe a variation is what's causing apparent variation. Some other feed. It's, it's the opposite. The feed itself can vary, can vary a lot. If you take distillers dried grains with soluble and you just look at it from one population, I know there are different processes right now, but you just don't know where it came from. That feed will be varying a lot from load to load and that variation will be real. And so that's why you better e either be sampling that quite quite often. And, and, but also we'll talk about that hopefully later on. There's other things that you can do from a formulation standpoint in a feed handling standpoint that will reduce the variation in the whole diet that's caused by feeds that are intrinsically variable.
Dr. Bill Weiss (00:18:34):
So if, if like I'll use soybean meal instead of cotton seed, you know, it, it's probably of the variation we see. Maybe 10% of it is true, the rest is apparent. What what does the user do then? I mean, do they, do they just use a book or do they sample?
Dr. Normand St-Pierre (00:18:50):
They just use a book. But here's what it is because in soybean meal you basically, and you have two families of soybean meal, you have the solvent extracted and then you have the, the expeller or oil process try to already have some that have the Hals in. And so you have the 48% soybean mill and the 44. Well, if you're buying soybean mill and you never test for it, I guarantee you sometime they'll stuck, stuck you with some 44% and make, make some good money on it. Not everybody's completely a hundred percent ethical in, in the, in the feed industry. So the reason why you should be sampling and, and, and feed companies do that sample and send every load to a laboratory, it's just be figured out whether it's been screwed up or not, that that's, that's all it's gonna tell them because otherwise, the the soybean meal, you're gonna have slight changes from year to year depending on the conditions of the crop. You may have a bit of regional differences, but these will be, it will be very, very minimal. The oil has been extracted, the hulls have been removed in, in solvent extracted soybean meal. And, and at that point the, the meal itself will be, will be quite constant. We'll just have a little bit of annual changes, but in the whole range of, of of diets, it, it probably is, is minimal. So that's why soybean meal, just use the table values and move on. Spend your money on something more useful.
Dr. Bill Weiss (00:20:16):
Yeah, one paper, I can't remember which one, but we, we kind of group feeds into where observer variation was much greater than than true variation. We just say use a book. But if say corn silage were were true and both true and and observer variation are quite high what, what's the, how, how can a user get to the true variation under practical conditions? They wanna eliminate this observer variation.
Dr. Normand St-Pierre (00:20:46):
Well, to, to, to, and, and the observer variation includes some things that happen in the laboratory. And I'm quite convinced that the commercial laboratories are overall quite good. That doesn't mean they're perfect. And so to, to be able to figure out how you're gonna apportion the variation, the variance is you have to repeat everything at least twice. So you're gonna have to take two independent samples on the same silage, but two independent samples. You're gonna send that to the lab and ask the lab to test them today and ask the lab again, test them two days from now and just, just run the same series of assay. Don't waste your time on minerals and things of that, that kind, these will have little effect there. If there, if the thing is to be able to separate the variation, look at the large analytical assay, crude protein NDF feed that contains some ether extracts probably could use that in, in dry matter.
Dr. Normand St-Pierre (00:21:52):
And if you keep track of those, you probably are nine winning the game 98% of the time. And so once you've done that, you should be able to partition then the variance. Now be careful. Standard deviations, even in things that are independent, are not additive. The variances are. And so for example, if if say the the standard deviation for observer is the variation caused by the observer is, is two percentage unit and the standard deviation from all the other factors that are the true variation is also two units combined. It will not be four, it will be, you add the variances. So the variance standard deviation is two, two square is four plus four is eight. Square root of eight, about 2.7, 2.8. So the total effect will be will be less than just if you had the two standard deviations.
Dr. Normand St-Pierre (00:22:52):
Those feeds that you're talking about, Bill, those that have about equal variation due to observer and, and due to true variation are those that you should really be spending more time on. Those that have true variation and little variances. Those basically, we know how to handle that through quality control charts. And you're gonna have to do repeated sampling. You manage to see if things have changed or not, and then you reformulate or not, or you can use base what's called in statistics or basin approach. You have a prior knowledge, you get a sample and you weigh the results based on the inverse of their variances. So this way, let's say your table value is 45%, NDF with a standard deviation of two and neural assay has come back at 50% and has also a standard deviation of two, they have the same variance.
Dr. Normand St-Pierre (00:23:54):
All you're gonna do is you're gonna average the two and take that average. If the variation is different, you just weigh them. And that's explain in the paper and that's really a pretty good approach. And so if you have actually, in fact if you have no clue about the variation they have in both, I've always said use the half step approach. Take what you thought it was before, take what you think it is today, average the two when you use that number, and next time you get a sample, keep averaging this way, you'll be only half wrong
Dr. Bill Weiss (00:24:25):
Dr. Normand St-Pierre (00:24:38):
Risk meaning minimize that and people are forgetting that there's two ways you can be wrong. The feed didn't change, the result did mm-hmm
Dr. Bill Weiss (00:25:43):
And another thing so if, if you make corn silage as, as substantial true variation, substantial observer variation, you take two independent samples, you send them in and then you just average 'em, the, the user would just average those numbers and use that number. Yep. And the, the question I get a lot is, okay, you did that and now you go back a week or two weeks later and you sample it. Should you then average that with the previous number? Or do you just say, no, this is the new number. How do you, how much averaging do we use?
Dr. Normand St-Pierre (00:26:17):
Well, there's two ways to handle that. There's the correct way, but it's more complicated. Use a quality control chart and you decide whether the change that you've had is significant enough based on the economics that never explained that in an old paper that nobody read. And so you can do that and, and that will be correct, but it's gonna get complicated. The other one is you just keep averaging
Dr. Bill Weiss (00:26:40):
So over the whole whole year or do you say, okay,
Dr. Normand St-Pierre (00:26:44):
Well when you get the new crop, I think by then there's a pretty good likelihood that things may have changed. At least it's a new crop and you go in with corn silage also alluded to that previously some of the analyticals will not change by the time you have a bunker silo and you, you load these up with a slope of 20 to 30% and it comes and, and nowadays we fail these silos so darn quickly and everything, but basically it, it's probably quite uniform in there, maybe not, but most time what it is, is you're gonna have a change through time. The dry matter will be changing and not necessarily because it's gonna rain on or something, but through the fermentation you're gonna have some change in, in the moisture content maybe one, one and a half percent. But definitely in terms of the availability of some of the matrices you have in the star, for example, we know quite well through fermentation through time the kernel will soften and the start will be made more available. Well, likewise, some of the stems, if you have in alfalfa silage through time also will allow microbial easier access to the stuff that's digestible within, within the the stems. And so you're gonna have change through time, but these changes are gonna be progressive. And so if you just keep averaging, you're gonna be following the progression of how things are changing, you won't be too far off.
Dr. Bill Weiss (00:28:05):
Yes, I, I disagree just a little bit. I think you can go too far. In other words, if I'm averaging date I had from January with date I have from July, I think you're, you're making things too smooth. And so I think you, you can go back too far, you know, feeds change, like you said. And if you're averaging the starch fermentation may not change hardly at all 'cause it's got fresh silage, itil. So I'd say I think well back a little bit from the whole thing, two or three.
Dr. Normand St-Pierre (00:28:34):
So what, and you have a valid point there that is if things will change by, by averaging you will be I said only half wrong, but you will always be half wrong and sometimes it'd be nice to be right in in which case what you probably should do is average, but a weighted average, the things that are further distance count a whole lot less. Yeah. And, and you can figure out some kind of weight that you would put in there. So the more recent results would be counting a whole lot more than older ones.
Dr. Bill Weiss (00:29:06):
Think the one more question on feed that's in this type, we have a table of co-varients or correlation among nutrients, I guess first again, in in English, not statistic ease. Why, why should users care about Covance and how should they use that, that information?
Dr. Normand St-Pierre (00:29:25):
The, the co-varients is used to calculate the correlation. The correlation means you have a linear association. If you have zero correlation doesn't mean two things are independent at all. It just means that they're not linearly related. Well, what does that mean practically? When you have a correlation, we think quite often, and I'll discuss this in terms of a normal distribution, it's say if you look and in the paper we give an example of corn silage and we use NDF and we use starch. And these two nutrients are highly negatively correlated. So if your silage average is 45%, NDF getting a silage with 47% is actually not very unusual. 17% of corn silages would actually have a higher NDF than that. So it's, it's kind of on the right hand side of the distribution, but not too far. It, it's it, and likewise, if the starch averages 35%, getting a starch of 38% is not that unusual.
Dr. Normand St-Pierre (00:30:34):
17% of sample would coming out with a starch at that level or higher. But getting a corn salvage with 48% NDF and 38% starch would be a highly, highly, highly unusual one in a one in thousand, maybe because the two are negatively correlated. The negative correlation means linear correlation. It means that when one goes up, the other one goes down and quite, quite, quite well that way. That also means that quite often if I just went in DF I'm gonna have pretty good idea with the star should be that silos. not perfect. I'm gonna have an idea if the NDF is high, I'm gonna guess that this, the, the start will be low. And so that's why it is important to look at these two because if you get a result and then suddenly those two nutrients that are negatively correlated make make them highly improbable that the, the these, these two results are actually real. And you should reach opponent and say another Singapore before you do anything else. It just means something goes cooky somewhere. And it does happen in a lab, you know, we run, we search lab and I'd say about I don't know about your Bill, but somewhere around between, depends on the grad student. Somewhere between two and, and 4% of sample have to be free assay because you got something that's absolutely goofy.
Dr. Bill Weiss (00:32:01):
I just see in the, in this paper there's a table and, and users could look at that to kind of get an idea of two, two feet, two nutrients are negatively correlated and one is really high and the other is both are high. You'd start saying, well, this might be a bad sample or bad data and, and get some new data or redo it. So it's, it's something useful. Just one more quality control check on these things.
Dr. Normand St-Pierre (00:32:24):
Yep.
Dr. Bill Weiss (00:32:26):
And then I, I guess I'd like to switch now to feed more, more feed, more diet stuff. And in, in this table I've given talks on this topic, I don't know how many times, and one thing I always made a point of is that multiple sources of nutrients reduces TMR variation in that nutrient. And we give a table here and, and people just have trouble with this. I I use corn size alfalfa and, and alfalfa NDF is more variable than corn side mdf on an average it's about almost 50% has a standard deviation, almost 50% greater. But if you use a blend of those two, you actually get less standard devi less variation in NDF than if you'd use all corn silage. So can you explain how that works?
Dr. Normand St-Pierre (00:33:16):
Well, I'll explain it first.
Dr. Normand St-Pierre (00:34:46):
And so now you understand since the variation of the hull, the contribution of a feed to the variation of the hull grows with the square, things that are quite variable, you feed less, if you just cut it by half, you will have reduced by four fold its contribution to the diet. And so if, if, if you think that the variation in corn sal is a factor half, two bunkers and just feed half from a bunker, half from the other bunker, you'll be reducing the variance of the whole, of the contribution of the corn sal the variance of the diet by four folds. And so that's why with your example, you can put alfalfa and corn salvage together. Now it's, it's 50 50 low 0.5, square 0.25, some of the two will be half of what you had before. And so that's, that's the, the beauty of mixtures now.
Dr. Normand St-Pierre (00:35:43):
So that's the mathematic. But if you just think of it, you, you would intuitively you would think mixtures ought to be much more uniform than any of the thing that you put in because the things that you put in, they, they they, they will vary. And if they're independent, which feed are no, the corn silage doesn't wake up one day and ask the pile of, of haulage, oh, how do you feel today? Are you hired an NDF? And then, then I will, now there they come, there's no correlation between the two. So if they're independent that way, then the errors will tend to self cancel. And the more that you have, the more they will be canceling each each one of them. So things that are quite variable, you use in smaller amounts, things that are not variable, you can put in larger amounts and anytime you mix things, if you they're properly mixed, you'll be doing a heck of a job at reducing the variation of the hole. That's, that's what a total mixed ration is supposed to be doing.
Dr. Bill Weiss (00:36:42):
And we, we put a, to me a a very useful example on this table with different ratio, the percent forwards in the D constant, we buried their ratio of corn size alfalfa. The alfalfa has a 30 or 40% greater standard deviation, and the optimum mix was about 40% or 30% alfalfa, 70% corn. It was much more consistent than the diet with a hundred percent corn silage over day-to-day variation. So again, these blends just are, are be if they're well made or just better, more consistent. Yeah,
Dr. Normand St-Pierre (00:37:20):
Yeah. Likewise, people buy some hay to which I say, well, if you think and you're feeding significant, if you feed one pound of hay, doesn't matter. But if you feeding, you know, you're out west and you're feeding significant amount of, of supposedly premium alfalfa hay. And if you think that variation is of concern, make two piles now two, and and if you're feeding 10 pounds, take five pounds from one and five pounds of the other one and you'll, you will be reducing significantly the variation to whole diet.
Dr. Bill Weiss (00:37:50):
Yeah. Again, you said they have to be made correctly. That's where the assumption occur. They're following the recipe, but mixtures or blends will be better than single, single sources of nutrients. Yes. The, the last part of this paper was talking about effects of variation on the cap. And, and speaking for myself, this was probably the most interesting research I did in my entire career. And, and I did a lot of research and the effect, the, the studies we designed to look at variation on on day-to-day variation on the cow to me was just remarkable. What, what can you explain, especially Peter Yoders, who was a master student for us. Can you explain that experiment a little bit, bit? And I thought that was again, just the, this is where he randomly picked a random pattern of forge NDF variation.
Dr. Normand St-Pierre (00:38:44):
Well, I think you'll be, you'll be better to explain that than, than I do because a lot, and and you ask where that came, came about, I ran an experiment and you know, those of you familiar with the Ohio State University, you know, we basically live in two different campuses in the same state, but may as well be across two different continents. So you have a group in Wooster, and we had the group in Columbus and we had a few cows and, and, and doing research on the Columbus campus was not easy. It's also a teaching herd and I'll spare you the details for the rest. But anyway, at one point I had a grad student, we had run this, this just neat little experiment. And basically they were, these were cows in mid lactation, about 35 kilograms of milk, 78 some pounds of milk.
Dr. Normand St-Pierre (00:39:30):
And, and we had three treatments. One of those, we fed the diet that was at 15% crude protein. And my golly, it, it, we did our best to maintain that consistent day-to-day basis, didn't change another diet. It was for one day they were on the on 13% and the other day on 17% and back on 13 and back on 17 and like this for three weeks. And the third treatment it was 11% and 19% and 11% and 19%. And so we really thought, especially the third treatment cows would out of bed. I mean, you just know when you get to 11%, this was the most boring set of results because no, normally you say, well, things were not statistically different. But here, when you go to the second decimal point and they had, they had no effect whatsoever. And so they just kept on 35 kilos of milk.
Dr. Normand St-Pierre (00:40:29):
There was no change in butter, fat test, no change in protein test. And so when you get something you weren't expecting, and you sit on this for a while, the grad student masters graduated, went somewhere else, and I went to Australia and forgot about all this. And, and, but it came kept hunting is thinking, well, if we had done it every two day at some point 11%, the cows, at least some of the cows would drop. We know that. But how long, what frequency for what nutrient? And this is where we started having beer together and wondering, well, for example, you, you told me that for vitamins, quite likely, at least for some vitamins, you could give the seven dose of vitamin in one day. So for six days you would have none. And we're all in one day. Anyone see hardly any difference for someone the nutrients, maybe the the frequency is daily by two days, and some others it might be a matter of hours, things that would affect luminal fermentation for example. And we didn't know we didn't know that. And so that's when I, because doing research in Columbus was so hard, I sort of passed the bat onto you and, and shipped Yoder to Wooster and said, Hey, you take care of it, that's, you take care of them. And that's when Peter ran, ran that experiment. And I'll let you discuss the results because you're more fluent with it since you, you were the one there with the cows
Dr. Bill Weiss (00:41:52):
Again. Peter picked a random pattern before the experiment ever started on varying forage NDF. And it ranged from like 20 to 29 or 27% of one of the exact numbers, day to day. Huge, huge vari more variation than we'd ever see on a farm. And these cows did absolutely nothing. They, compared to the control was as consistent as we could do and and more consistent than you'd ever see on a farm. And this highly variable one, which again in a random pattern would high, low, low, low, high, all over the place and production was about a hundred pounds on, on both treatments. The only con the, the limit here was both treatments on average over a 21 day period were, were identical. They both were the exact same forge, NDF. And when I talk about this, this, people would say, well, this means it doesn't matter what we feed a cow.
Dr. Bill Weiss (00:42:47):
No, what it means is you don't, it doesn't matter every single day what you feed a cow. And, and that to me is an important question, and you brought this up, is how long can I feed a a bad diet before the cow will respond? Because that would really be important to dictating sampling frequency and reformulation frequency. And like I said, vitamins and most minerals you could sample monthly or even more maybe or even less frequently. But things like starch, you know, they may respond 'cause that'd be a ruminal effect. They may respond in a few days who I don't know. Protein, again, you you were talking, we did a study where we also fed 11% and after about three days things started going bad, but not as bad as you think or not as fast as you think. And it makes sense. 'cause You know, every day we don't start with an empty room there, there's a lot of stuff still in there. So if it's low the next day, you know, a third of the diet from the previous day is still there. So the cow buffers a lot of these day-to-day variations as well. So
Dr. Normand St-Pierre (00:43:54):
Yeah. And I, I like to remind people that although we've been breeding cows for a few hundred years now, they carry genes that have been selected over millions of years. And there is no way a ruminant would survive in the wild if every day or every hour they needed to have something that is quote unquote balanced. Now mind you, those ruminants don't produce 25 or 35,000 pounds of milk. So that might shorten it might reduce somewhat the adaptability of the animals, but I'm convinced that you have have some, these genes and cows to to cope with that in our experiment where we're varying the, the protein, the milk here in nitrogen was tracking exactly with the diet. I mean, you could just see that plummeting and jumping right up. And sometimes we forget that this urine nitrogen diffuse across all the, the soft tissues.
Dr. Normand St-Pierre (00:44:52):
And so when there are 19% crude Putin diet, you have an increase in urine nitrogen in all the tissues, and then you put them for one diet root, some of that gets reabsorbed. And so cows are not, are not, you know, I've seen pictures where they're compared like a Formula one car where, you know, a little tidbit not being, then they're gonna crash. And, and that's not, that's not, we crash them. I'm, I'm afraid to say this, but I think as nutritionists, we crash the cows more often than it would crash otherwise because we changed things that shouldn't, shouldn't have been changed.
Dr. Bill Weiss (00:45:31):
I, I'd like, again, with this Peter Yoder study, people say, well, it doesn't matter, but, but DMI dry matter intake tracked forge, NDF, like everybody says it should, but again, it never was that high or that low for that long. These cows didn't bounce back, but it, cows do what they said. And if we would've kept NDFs really high for four or five days, if that was the rampant pattern, we would've saw something. But if it's high for a day or two, then they just make up the intake when it gets lower, when they'll forge NDF, they eat more and go right back.
Dr. Bill Weiss (00:46:05):
The towns responded, but they just, it wasn't bad enough, long enough.
Dr. Clay Zimmerman (00:46:09):
So short, short term, I, I mean they have protein energy reserves to make up for that, right? Yep,
Dr. Bill Weiss (00:46:16):
Yep.
Dr. Clay Zimmerman (00:46:17):
So what, what about the dry matter variation in these forages and, and the study you ran with that, I was pretty intrigued by that. Yeah,
Dr. Bill Weiss (00:46:25):
It is. The one is for dry, dry matter silages are more variable than the other nutrients because of brains. So there is a, a greater true, true variation dry matter than the other ones. But Lucian Macbeth and other master student did a really neat study where he, he made silage wet like it rained and he increased the, or decreased the dry matter of these silages by 10 units. So it was a lot of rain and he put, and so and these, these experiments last three weeks, so twice during this three week period for, for two three day periods, he fed this wet sage, nothing else changed, just he just added water to it and fed it and, and then with control obviously would never change. And what, what happened is, and, and we always made sure feed was available, so when it was wet, we added more as fed feet.
Dr. Bill Weiss (00:47:21):
So the, the bunk never ran out. And that's the key thing here. And what happened the first day we fed the wet stuff as fed intake went up, but not as much to account for. So DMI dropped, but as fed intake went up by the second day as fed intake went up even more and dry matter intake was the same as the control took them one day. And milk dropped during that and which we kind of expected, but then when we put 'em abruptly back on the control, as fed intake stayed high for a day, so they actually dry matter intake increased the day following the wet silage and it stayed high for two days. So they made all it back. You know, the, on average there was no effect in intake, no effect in dry matter, but it dropped when it was really wet.
Dr. Bill Weiss (00:48:13):
But then when we went back to the dryer stuff, the normal stuff intake increased and it, it's like these cows that took a few days for them to figure out I have to eat more and then no, now I have to eat less to maintain intake. And, and it's always intrigued me if we could use this as a management tool, like once a week or something, you feed really wet feet, wet sage and then go right back and see will they, will that bump following the, the change in dry matter? Well that can be consistent and actually end up with more milk. So
Dr. Normand St-Pierre (00:48:48):
Now Bill, there's a very important point here that was important in that experiment is the total amount fed was not limiting. That is there was always feed there. Exactly. And so from that, my, my hypothesis is really what we see in herds that crash when, when it's because the feed, the silage gets wet, they feed the same amount as is, and so you have less dry matter, the cows run out of feed for four or five hours and that doesn't work that we know. And so when you have bare bunks, and so if the solid is wetter, just feed more of it, don't change the rag, the the formula itself just feed more of it. Yeah. And so it's a matter of bunk management as opposed dry matter management because I'm convinced myself that's a good example also, if it's wetter today, when is it not gonna be so much wetter? And, and at some point you, it's like a dog running after it's stale. And, and as long as Calgary have sufficient feed that they don't run out in the, in the, in the feed bunk, I think they do fine.
Dr. Bill Weiss (00:49:54):
Yeah, we, in that study we also, there was a treatment where we rebalanced it to account for the dry change that was actually worse than when we just fed more the, the, the original TMR Yeah, it was, we made no adjustments just feed more.
Dr. Normand St-Pierre (00:50:11):
Yeah, it was counterintuitive, but it was very clear, I mean very clear.
Dr. Bill Weiss (00:50:15):
And there's a study like I think it was outta USDA for Dave Mertons where he, he did the same thing, but he, he didn't feed more feed cows ran out and those feeds did, those cows did substantially worse. So if you have to make sure increased delivery rates of the TMR, but I wouldn't get too excited about changing the forge concentrate as the, as the silage got wet.
Dr. Bill Weiss (00:50:43):
I guess one, one other thing we did in this, and again, these types of papers they live, you let authors speculate a bit. We spent a a big section on where, where do we think we should go, what research is needed and it'll, it'll be conducted by not you and me anymore, but, So what, which you Norman, if you had to prioritize, what would be the, the, you got a young professor here, which, which project would you tell him to do of the we gave five or six what we thought would, should be done? Well, I
Dr. Normand St-Pierre (00:51:12):
Think first I would advise him or her to go and take another line of work because getting, getting funded to do these type of projects will be difficult 'cause it's not sexy. He doesn't have genomics, it doesn't have neutrinos and it doesn't have ome no more in there. And so I don't know the, the, the one thing that really needs to be clarified is basically this, this period, how, how long the amplitude and, and the phase, how long cows will take to respond to a change in the major nutrients. And so we're talking here of rum degradable protein maybe metabolizable protein, although that's calculated number, but certainly NDF and forage and df and really change the amplitude of the changes in the length of time that, that it does happen to calibrate to see how how that's gonna change. So that's probably the one that for me would be the most important in terms of the economic impact on the herdz and the economic impact in monitoring the nutrition in those herds.
Dr. Normand St-Pierre (00:52:23):
We spend an ordinate amount of money doing analysis, analysis, analysis of feeds. And I think that some of it, it's misplaced. There's some feed that we should be doing more analysis and some nutrients, more analysis and some of the feed we should completely forget about doing any kind of analysis and just move on with our life. But we really don't understand very well the cow buffer on that. The other thing that is, we, we didn't put in that paper, but it's kind of embedded in there is, is we never feed a cow. We feed a pan of cow. And so in that pen you have cows at 80 pounds of milk and you have a cow at hundred 50 pounds of milk. So a pan doesn't have a requirement. So when we talk, we, when people say we feed, if the, the pan averages a hundred pounds of milk, it means that if you feed 400 pounds, you're gonna be shortchanging half of the cows.
Dr. Normand St-Pierre (00:53:20):
And so that's leads to some of the lead factor that have been and projected. The the point being is when you feed a pan, you no longer have requirements. You have responses to nutrients. And so depending on where you're at in the level compared to the average in the pan, the pan might be more or less susceptible to nutritional variation. If you feed a very high plane of nutrition, you're gonna be spending more money feeding those animals, but you will be buffered because you're always kind of over supplementing. If you wanna think that way. Energy's a little bit different for that. And so that, that's also an aspect that is very not often studied. That is what's, for example, when we did these, these experiment and, and we're looking at how to, again, we're just look at on an average, well if nothing happens, probably not nothing happened to any cows, but when you start having some production responses due to, to variation is in the high producers that trashes or the medium or the low, I think my hypothesis would be it's the highest one you'd knock them on the head, but that hasn't been done.
Dr. Normand St-Pierre (00:54:30):
And and a way to do that would be to have factorial approach where you have cows of different potential different looking ability and look how they respond differently to variation.
Dr. Clay Zimmerman (00:54:43):
Did so it, in, in, in the the studies that you've done, you know, looking at, at the feed variation, were there really fresh cows in those studies or were these cows a little further into lactation?
Dr. Bill Weiss (00:54:57):
All these were, some of 'em were early lactation, but not fresh, maybe two months. And again, we were afraid of what would gonna happen to these cows because some of this variation was huge, so we didn't want fresh cows. But that's a a good point. Is it, are they more susceptible? We all think they are, but maybe not. But that, that, to my knowledge, no one's done any of this stuff with, with fresh cows or even real early.
Dr. Normand St-Pierre (00:55:24):
Yeah. And, and so to summarize my answer to your question, Bill, there's some stuff that has been looked at because it was relatively easy and cheap. That is to partition, for example, the variation in feed between the, the, the observer variation, the true variation, all of that compared to an animal trial is a whole lot simpler and a whole lot cheaper. There's just not a whole lot of animal experiments. And multifactor, again, like I said, to look at what animal factors might modulate the response that they will have to variation in the magnitude, the amplitude of the variation in the duration of death, duration of different phases that becomes more onerous. Also imagine if you start doing experiment where you think that you're gonna be screwing up cows in a lactation, you think that is very popular in your department.
Dr. Normand St-Pierre (00:56:21):
So at universities you're gonna have some problems there. In commercial herds you have an n of one, you have one fresh pan. So it's, it's very, I've during my, my, my career I was visited many, many times about people who were doing commercial experiments in herd. And I probably was saying, oh, the sick, the sick experiments because they were coming to me when things didn't go out as, as it was planned out a lot of time I could tell them just what it died of. But you know, replication is pretty fundamental in research. You don't have replication. I can't tell.
Dr. Bill Weiss (00:56:58):
Yeah. I also say, you know, these fresh pans, the, the cow variation is gonna be a whole lot higher because you know, you're gonna cow have cows that one day of milk eating hardly anything. And cows, that three or four weeks, they're eating, you know, 50, 60 pounds. So that would even make it more com Maybe it makes it less, the pen less responsive to variation because, 'cause the cow variation is so high. So that'd be really, really tough experiments to do
Dr. Normand St-Pierre (00:57:25):
So. Yep.
Dr. Bill Weiss (00:57:27):
And guess before, before we end here, I, we do wanna put a plug in. You said funding is a major issue. USDA did fund a lot of this research. So do, do some practical research,
Dr. Normand St-Pierre (00:57:39):
I guess. Well, I think what, what it was Bill though is, is the way we Yeah. The way we worded that out. Oh yeah. And so and you were, you were, we were really good in there at, at writing those grants. I would not, I would not have got the fund, I didn't have the right verbiage. You have to put in a ways that the people reviewing understand, but also in a way that you people won't understand the impact. It can, it can have. And so we did a masterful job of doing that. So USDA, you can, you couldn't go and get funding, but it'll be for a young faculty, it'll be a whole lot easier to get tenure
Dr. Bill Weiss (00:58:23):
Yes. You know, I asked you the question, what would you prioritize? To me, what I do is this, this idea of, I call controlled variation changing, say, change the, the ratio of corn silage and alfalfa once a once a week or something, will, will this stimulate intake? You know, we, we, it works for humans, it works for pets, it works for zoo animals. So this would be an area I wish people would look at. Can we every so often change something and will the cows respond positively to that change? I think that's an intriguing area of research.
Dr. Clay Zimmerman (00:59:01):
So, so Bill, I, I, yeah, I was curious about that. I think that was the last question that you had, that you all had in the paper. So you, and you talked about zoo and lab animals. So what, explain what's happening there?
Dr. Bill Weiss (00:59:15):
Well, like, well, for a zoo animal, they might put in new a different fruit. Say they're, they're, they're vegetarians and they eat on Wednesday, they feed oranges. And then all of a sudden they just behave better. They, they may eat more, they do better. And so will, will this diversity and diets Well that, you know, we all, I I can remember from the first nutrition class I ever had being beaten to my head. Cows love to be bored. They kept saying this and it, no one ever proved that. And so maybe cows like a little variety in their diet and maybe they, they'll respond positively if all of a sudden I throw some hay in a diet that's never had hay before, they say, boy, this is, this is good. I'm gonna eat more. And then they're gonna produce more. So I think stuff like that, that the dogma very often is wrong. And because, but no one ever tested. So I, this diversity in diets I think would be an intriguing, whether it works, who knows? But it would be an intriguing area of research.
Dr. Normand St-Pierre (01:00:19):
And, and that's a good point because it, it, it reminds me of a story in my, in my life at one point I was involved with people in animal behavior and they have the five freedoms that they want to have for animals. And one of those is animals being stress free. And I said, I always said, well, if the five freedoms apply to the animal should apply to humans as well. So humans should have stress-free life, but yet on, on, without being forced, you have humans going and running marathon. That's a little stressful if you've done one sucker like that. And, and we know when we were in grad school and we had to go and defend our thesis. Was that stressful? Yeah, but you passed in, man, you felt really very good after that. And so I, I've always thought that the rabbit needs a little stress. He's pretty happy when he, he's escaped the, the, the wolf that was running after him or something. And, and maybe the cows need, need something to get them out of that boredom. I, I don't know. But like you said, this idea that cows need to be put in complete boredom have never been tested. And I think animals in general probably need a little bit of stress. They don't need a chronic stress, the same stress all the time. A little peak here and there. At least in my life. It, it works well for me.
Scott Sorrell (01:01:38):
Well, gentlemen, I've been a, a spectator today and, and gladly. So this has been fascinating. I've enjoyed watching the conversation. As we get ready to close here Bill and Norm and, and, and Clay, give us kind of a few final thoughts. You know, what are some key takeaways that you think our audience ought to take from this? And, and Clay, I'd like to start with you if you don't mind.
Balchem (01:02:09):
New research is changing everything we thought we knew about Cho's impact on the cow and her calf and top scientists have a lot to say about it. They're presenting new research that supports choline as a required nutrient to optimize milk production choline as a required nutrient to support a healthy transition choline as a required nutrient to improve calf health and growth and choline as a required nutrient to increase colostrum quantity. This new research is solidifying cho's role as a required nutrient for essentially every cow, regardless of health status, milk production level, or body condition score. Learn more about the science that is changing the game and the choline source that is making it happen. Reassure precision release choline from biochem. Visit bal chem.com/scientists say to learn more.
Dr. Clay Zimmerman (01:03:11):
So I really love this paper. I've never read a paper quite like this. I'm would encourage everybody to read this. I liked how you wrote, there's full roles that they wrote in the paper that we've kind of covered during the conversation and then it ended with the seven research questions that that Norman and Bill, they hit on two of those. But I I I think it's a fascinating paper and, and a really good read. So I, I would encourage everybody to go read that there. I think there are a lot of, a lot of take homes here related to, to sampling or on the flip side, you know, things that you really shouldn't bother with sampling it and using the, the book values that are out there.
Scott Sorrell (01:04:02):
Yeah. Good comments. Clay, norm, do you have any final thoughts for us?
Dr. Normand St-Pierre (01:04:07):
Yep. If you're gonna be sampling or feed, do it twice minimum. Otherwise, otherwise you have no idea about variation. Two, try as much as you can to separate the observer variation from the true variation so that you don't end up like a dog running after it's stale sample, which would be sample a nor Raf. There is some feed you may better might as well just use the book value and, and also concentrate on a few critical nutrients as opposed to every time you see in a sample, you run the $90 assay with everything with the kitchen sink in there. If something changed normally you would see it in some of the main indicators and for assays that have more repeatability than some of the biological base ones.
Scott Sorrell (01:04:56):
Alright Bill, we're gonna let you bring it home.
Dr. Bill Weiss (01:04:59):
Well, again, I just encourage nutritionists to think when they get new data, to sit down and think before they get it, their computer and just punch in a change and, and ask I'd was beating my students' head as they get new data. I ask them, why do you think it changed? And if they said, I don't know, then I'd say, let's just leave it. Go and, and go. So think when you get new data, think before you change and ask yourself, do you think this is real? Don't, don't be afraid to use your, your brain on, on changing things or not changing things.
Scott Sorrell (01:05:33):
Yeah. Well, thank you gentlemen. Thank you for your time, your knowledge, for spending time with us here. It's been a great conversation. Really do appreciate it. To our loyal listeners thank you for joining us once again. I hope you learned something. I hope you had some fun and I hope to see you next time here. It's the real science exchange where it's always happy hour and you're always among friends.
Balchem (01:05:56):
We'd love to hear your comments or ideas for topics and guests. So please reach out via email at anh.marketing at balchem.com with any suggestions and we'll work hard to add them to the schedule. Don't forget to leave a five star rating on your way out. You can request your Real Science Exchange t-shirt in just a few easy steps, just like or subscribe to the Real Science Exchange. And send us a screenshot along with your address and t-shirt size to anhmarketing@balchem.com. Balchem’s Real Science Lecture series of webinars takes place on the first Tuesday of every month with the top research and nutrition topics that will impact your business. We also include small ruminant, monogastric, and companion animal focused topics throughout the year. Visit balchem.com/realscience to see the upcoming topics and to register for future webinars. You can also access past webinars and search for the topics most important to you.