I hope you'll find this paper useful: http://oceanrep.geomar.de/3643/1/317_Froese_2006_CubeLawConditionFactorAnd_Artzeit_pubid8605.pdf

WZ

You could try analysis of covariance (ANCOVA) of fish weight with length as a covariate. It works well and avoids the need to use ratios (such as the K).

In addition dry weight rather than wet weight is more accurate since fish tend to compensate loss in weight (for example loss of protein) with water.

The answer you can find also in this paper: J. E. Williams. 2000. The Coefficient of Condition of Fish. Chapter 13 in Schneider, James C. (ed.) 2000. Manual of fisheries survey methods II: with periodic updates. Michigan Department of Natural Resources, Fisheries Special Report 25, Ann Arbor. (http://www.michigandnr.com/publications/pdfs/ifr/manual/smii%20chapter13.pdf)

I think one of the major limitations is the frequent over-interpretation of this simple numeric representation of a length-weight relationship.

Firstly, by calling this ratio a condition index or condition factor leads many authors to draw conclusions that one group of fish with higher K was in a better physical condition than another group with a lower K. If we would make the same assumptions for human, we'd say that those individuals or groups with the highest BMI had the best condition!

Secondly, many authors draw conclusions without understanding that in fish as in other vertebrates, linear length growth which is primarily skeletal growth is a fundamentally different growth process from weight growth, which can be due to protein accretion and increased muscle mass, but also due to fat deposition or gonadal growth. Length and weight growth are regulated by different endocrine systems and may not necessarily happen at the same time. These growth processes can affect the fitness of the fish in many different ways and have different energetic costs.

An example of this is when anadromous salmonids go though parr-smolt transformation. There's a very significant decrease in K. This does not mean that the smolt is in a worse physical condition than the parr, The K decreases because the fish are growing rapidly in length relative to weight, changing their body shape to a longer, leaner shape. Ultimately, this is advantageous for their transition to the marine environment where they have to swim for long distances in open water.

Hi Bjorn, thanks for the insight into some ecologically relevant reasons behind certain physiological changes in fish, but i'm really curious to know your own point of view on the possible options and why?

Your question was actually about whether K can be used "as a measure of fish health" and I'd say No! A condition factor may or may not be affected by any number of bacterial, viral, fungal or parasitic diseases, pollutants or other conditions which affect the HEALTH of the fish. In many cases, the fish probably die before the CF is affected. So, if you really want to study the health of the fish, you need to look at health-related parameters, e.g. immunological parameters. I'm not a fish pathologist, so I hope some expert in the field can add relevant comments to this.

Under controlled conditions, it's possible to use CF as a measure on nutritional status, but I wouldn't call that "health".

Another general aspect of using ratios is that you've probably already carried out statistical analysis on the primary data (weight and length) and then you calculate a ratio, and then also analyze this statistically. In reality, you've then tested the same material twice for statistical differences, increasing the chance of false positives.

From what i infer from your response Bjorn, nutritional status or food conversion efficiency otherwise called condition factor of a fish does not necessarily imply a healthy fish and should on safe grounds be limited to controlled set-ups. Thus the word 'health' should not be used in place of fish condition factor.

I would say this is an important clarification. Thanks a lot.

The biggest problem with Fulton's condition is the assumption that fish are growing isometrically, hence the exponent on length is assumed to be 3. Better approaches include relative condition factor Kn = W/a L^b where a and b are constants determined for the population from empirical data or Relative weight Wr = W/Ws x 100 and Ws = a' L^b' where a' and b' ideally account for the genetically determined shape characteristics of a species. When Wr is significantly greater or less that 100 it means that fish are heavier of less heavy than the "standard". Again, this assumes that a' and b' are known. Personally, I think the slope of the linearized Length -Weight regression relationship is the most useful measure.

It's possible to use CF as a measure on nutritional status, but only if comparisons are made ​​between groups of individuals of similar size. This is because in most cases the fish have an allometric growth, with a regression coefficient length-weight different from 3. If there is allometry, the value of FC is conditioned by size. To eliminate the effect of size I recommend to use the relative weight, which is an index of condition significantly better than FC. In this regard, you can read: Blackwell, B. G., Brown, M. L. & Willis, D. W. (2000). Relative weight (Wr) status and current use in fisheries assessment and management. Reviews in Fisheries Science 8, 1–44.

Personally, like Y. Afeworki, I'm more of a fan of using ANCOVA[1] to compare relative mass for a given length, but the assumption of a common slope can cause problems. The alternative (which gets criticised by some[2], but used by many[3]) is to run a single mass-length regression and then compare residuals.

If we really want t understand condition, I think we have to look at lipid content. Does anyone know of papers directly comparing condition (e.g. K) with lipid content?

[1] Trippel, E.A. & Hubert, J.J. (1990) Common statistical errors in fishery research. Writing for Fishery Journals (ed. J. Hunter), pp. 93-101. American Fisheries Society, Bethesda.

[2]García-Berthou, E. (2001) On the misuse of residuals in ecology: testing regression residuals vs. the analysis of covariance. Journal of Animal Ecology, 70, 708-711.

[3]Jakob, E.M., Marshall, S.D. & Uetz, G.W. (1996) Estimating fitness: a comparison of body condition indices. Oikos, 77, 61-67.

I think that Fulton -k index is very useful even in modern conditions of fish farming,for rutine work.(estimate health status of fish for winter period).Fish affected by parasites,bacteria,virus or bad water factors stop eating and F-k index refflect it. You may use modificated index:not full length but scaled body's length.Also it's possible to use Clark index: in the same formule to take the weigth without weight of ovary and intestine.

The great advantage of the index of condition and of the relative weight is they do not require the sacrifice of the fish, unlike other methods.

Hi, Jaroslav, I've really been eager to read the paper you recommended, but it seems there's quite some issues with the link. Could you kindly help out, i really appreciate your contribution. Thanks

I can certainly agree with Rita's and Massimo's comments on the usefulness of the condition factor when it's used in a correct way, and I most always use this as one of the descriptive parameters in my research on the endocrinology of growth and appetite in fish.

However, as the question concerned the LIMITATIONS of using it, I highlighted the common overinterpretation of this parameter in my original response to the question.

Generically Fish condition is described as the well-being or robustness of an individual fish

Measures of condition are intended to be an indicator of tissue energy reserves, with the expectation that a fish in relatively good condition should demonstrate higher growth rates, greater reproductive potential and higher survival than a lower conditioned counterpart, given comparable environmental conditions.

Fultons condition factor assumes isometric growth (b = 3) or growth without changes in body proportions, resulting in condition factors that are often length- and species-dependent. However, body form changes with length (b > 3) and species, thus K increases with increasing length. This limits its application to fish of similar length within the same species.

Relative condition factor according to Le Cren (1951) attempted to solve the deficiencies of Fulton’s by comparing the actual weight to a standard weight predicted by the weight-length regression based on the population from which the fish was sampled.

Dear Azubuike Victor Chukwuka and colleagues,

Another relevant aspect to take into account is that morphometric condition indices take longer to show the effects of food intake (Catalán, I., Berdalet, E., Olivar, M.P., Roldán, C., 2007. Response of muscle-based biochemical condition indices to short-term variations in food availability in post- flexion reared sea bass Dicentrarchus labrax (L.) larvae. Journal of Fish Biology 70, 391e405.7).

In this other paper we compared the Fulton index with the TAG/CHOL ratio as indicators of fish larvae nutritional condition.

https://www.researchgate.net/publication/234155609_Ecological_understanding_for_fishery_management_Condition_and_growth_of_anchovy_late_larvae_during_different_seasons_in_the_Northwestern_Mediterranean?ev=prf_pub

Article Ecological understanding for fishery management: Condition a...

Dear David, thanks for your contribution, but i think measures of morphological condition is easy relatable to ecological circumstances compared to biochemical condition indices. Better still, i recommend a complementary usage of both indices.

Hi Azubuike and colleagues-

This is an interesting question that is broader than just fish, although my experience is with fish and crayfish. I agree that we have come across situations where the index has been wrongfully applied in our reviews and experience. My opinion is that operator error can suspected, but likely only accounted for if one has access to the original information? To verify the reliability of the data is indeed important, but generally there is no way to check this with the exception of identifying low values. Our experience is that b-values approaching zero would give a strong hint that something is wrong. Values should range between 1+ and 4, typically.  

A very good paper is,

Stevenson RD, Woods WA Jr. (2006). Condition indices for conservation: new uses for evolving tools. Integrative and Comparative Biology 46: 1169-1190.

This paper compares the variety of body condition indices and puts into context of other zoological studies. My lab and collaborators believe that the Fulton Condition Index and other body condition indices have not been explored sufficiently.

We are using this index extensively in the development of both species specific indicators, such as the relative weight indices and also for comparison of ecological interaction predictions for alien introductions.  I think that when we consider the relationship of length-weight, allometric-isometric growth rates, and compare to body morphometric patterns that this has very important "basic" life history ramifications. We are exploring ways to look at these condition indicators as a "basic tool" or indicator of population stability especially when comparing to changing landscapes conditions.

Best wishes,

Thomas P. Simon, PhD

I think this paper is very logically reviewed this issues and projected the limitations with a good comparative analysis. You can go through the history section of the paper enclosed herewith. Thanks.