Mario Ernst wrote with this interesting question:

I have been in contact with a lab in the UK for the past few days discussing the suitability and possible limitations of Mehlich 3 with regard to MLSN.

Below is an email I received today and I am very interested in your thoughts on this. If you have time you could post your view on this on your blog as I think this could be of interest to anybody thinking about using the MLSN guidelines.

Here it is: 

[from the laboratory]

Hello Mario,

Yes, you are absolutely right in your thinking that Mehlich III (M3) has gained popularity recently due to its ability to provide information on a great number of nutrients with a single extraction. Therefore, its costs are low and turnaround rapid.

However, it does have, in my opinion, some drawbacks.
Firstly, routinely only 2ml of soil is used. A very small amount, i am sure you'll agree. This 2ml, therefore, has to representative of all present nutrition. Be it a major, secondary or trace nutrient.

Secondly, M3 will over extract cation content compared to traditional exchangeable cation techniques. Especially calcium and magnesium on calcareous soils. This explains the lower guidelines in the MLSN system. It's not wholely due to a change of agronomic thinking but a reflection of M3's more aggressive nature.

Thirdly, as M3 is a weakly acidic extraction, it can quickly be neutralised by calcareous soils resulting in its ability to extract trace nutrients plateauing. I routinely see very low metals contents on calcareous soils tested by M3.

Having said all that, M3 most certainly has its place in modern agronomy and we do perform this test upon specific request. But, in addition to my observations above, as with any system, it is only as good as the guideline system employed for result interpretation and the trials work conducted to produce those guidelines.

This is why our preferred option, and that of all our major customers, remains to perform bespoke testing as per the details in our methods statement. 

Those are interesting points, and ones that I do have some comments on.

a. Mehlich 3 is the extractant with the most widespread use in the United States. For the development of the MLSN guidelines, we didn’t choose Mehlich 3 because we think it is the best extractant. Rather, we chose it because it is used at soil testing laboratories across the United States, including at some of the labs with the highest volumes of turf samples, and it is as much of a “standard” universal extractant as there is today. With Mehlich 3, we have access to a lot of data, and the guidelines produced for Mehlich 3 extractions will have the broadest application at the greatest number of turfgrass sites.

Because Mehlich 3 is so widely used, there are equations available to estimate what results would be on other common tests, and vice versa. I will share some of that research on this blog in the future.

I would prefer that all turfgrass soils were tested with 0.01 M SrCl₂, which I think has a number of advantages over Mehlich 3 as a universal extractant for turfgrass soils. But since that is not realistic at this point, I do a lot of work with Mehlich 3 data.

b. Yes, the standard procedure for Mehlich 3 and for many other laboratory is to use a small volume or mass of soil, usually 2.5 cm³ soil or sometimes 2 g of soil. I don’t see that this is a problem. Soil samples submitted to a laboratory are dried and ground to pass a sieve. The sample to be analyzed is drawn from the prepared soil, and that sample is representative of the larger sample. I’ve not had a problem with this. For example, I applied K to a research putting green and then measured the amount of K in the soil. The amount measured by soil tests, even though the test analyzes just a small amount of soil, is representative of the amount actually in the soil.

c. Mehlich 3 is never a suitable extractant for calcium in calcareous soils, because the acidic extractant dissolves calcium carbonate. And depending on the presence or not of magnesium carbonate in the soil, Mehlich 3 may or may not be suitable. But this doesn’t really matter, because calcium won’t ever be required as fertilizer in a calcareous soil. One doesn’t need to test for it. So the erroneous calcium data can be thrown away. And the same goes for magnesium. If there is magnesium carbonate present, and the extractant dissolves some of it and gives an erroneous result for magnesium, it doesn’t matter because magnesium was not required as fertilizer in that situation.

d. As for why the MLSN guidelines are lower than conventional guidelines, it isn’t because of Mehlich 3 over-extracting cations compared to other extraction methods. In fact, the MLSN guidelines calculated from Mehlich 3 data are higher than what one would get if the same soils were analyzed with extractants that removed fewer cations than Mehlich 3.

e. The Mehlich 3 extractant starts at a pH of 2.5 and after mixing with soil the pH will change, and unless one measures it, one won’t know what the pH of the extraction was. That is one of the reasons I prefer 0.01 M SrCl₂ — because the extractant adjusts to soil pH and we know that the extraction was done at the same pH as the bulk soil pH.

f. As far as micronutrients with Mehlich 3 in calcareous soils, I have a few thoughts on this. First, we have deliberately avoided micronutrients in the MLSN guidelines, because we have not studied this in detail. We may study this in the future and propose some MLSN guidelines for micronutrients.

There may be cases when Mehlich 3 is unsuitable for micronutrients. But micronutrient testing of soils is fraught with problems anyway. And turfgrass micronutrient deficiencies are so rare as to be almost unseen. In the case of moderate to high pH soils, one would generally apply micronutrients anyway, to make sure there is availability, and would not consult soil test results. In fact, I rarely look at micronutrient results, and I’m not aware of good guideline levels for turfgrass no matter what extraction method is used. There has not been much research on this topic. So I am not defending the micronutrient data from Mehlich 3, but I am saying that I don’t think they are especially useful as decision-making criteria for turfgrass managers making fertilizer decisions, nor are micronutrient data from any other method.

g. Regarding “it is only as good as the guideline system employed for result interpretation and the trials work conducted to produce those guidelines.” I agree completely, and I think that is the most important thing to consider when doing testing. Just saying that, however, doesn’t convince me that the lab in the UK has a better guideline system for result interpretation than the MLSN guidelines. As I’ve explained, conventional guidelines are high, broken, and a main reason for development of the MLSN guidelines was to have a set of guidelines that is representative of the soils in which turfgrass is grown today.

h. Bespoke testing with a number of different tests depending on the element or the soil would have some advantages. But again, it comes down to the guideline system used. With the MLSN guidelines, we are working to be open about what data we are working with and how we generate the guidelines. For those who prefer to use other guidelines, I would like to see why those levels are what they are, and just how extensive and convincing is the trials work done to justify those guidelines.

From what I’ve learned about conventional guidelines, and this comes right from the textbook, they were in many cases set artificially high, based not on agronomic research, but because it was considered that cost of fertilizer was not important in turfgrass management.