THE JACKSON RATIO- A MEASURED VIEW (letter in reply to a query)

Dear Chris

Thanks for your interest in our use of the derivation from Jackson's graph as
a mathatical ratio.

I have spoken to Robin Bone, who first introduced us to this method, but his
research is unpublished, and typically for him, lost at this time. He can site
however, work from other researchers who back up his adoption of the wt (g)
divided by the length cubed (cm) as an accurate formula which replaces and indeed
accurately matches Jackson's graph.

In the Veterinary Record (1995) 136, 566-568 Blakey and Kirkwood explore this
avenue with RETs, box tortoises, Hermann's and Spurs. Paragraph five quotes
them as follows: "between bodies of constant shape volume is related to
the cube of linear dimensions. Amonst these animals weight was approximately
proportional to length cubed...."

Not having read this five years ago, I simply took a practical physics approach
to this, and put the Jackson graph to some simple tests:

I took 12 points on the graph along the "average line", and calculated
the weights and lengths, based first on the weight divided by length, which
is widely quoted as "The Jackson Ratio", and then using wt div by
L cubed, as promoted by Robin Bone.

The results to me (a lapsed physics teacher) were startling:

weight(g) | Length(mm) | wt(g)/L(mm) | wt(g)/L3(cm) |

300 | 110 | 2.72 | 0.22 |

400 | 120 | 3.33 | 0.23 |

600 | 136 | 4.41 | 0.24 |

800 | 149 | 5.33 | 0.23 |

1100 | 168 | 6.54 | 0.23 |

1300 | 180 | 7.2 | 0.22 |

1500 | 111 | 7.8 | 0.21 |

1650 | 200 | 8.25 | 0.20 |

1900 | 214 | 8.87 | 0.19 |

2200 | 230 | 9.6 | 0.18 |

2400 | 240 | 10.00 | 0.17 |

2500 | 256 | 10.1 | 0.17 |

My mathematical friends are quite clear on this issue. Jackson's graph is for
most of its length a straight line, and therefore can be expressed as a single
ratio- i.e. a formula which when mathematically calculated, defines any point
on the graph with the same resulting number. It is clear that the simple calculation
of weight divided by length does not meet the definition of a mathematical ratio.

The weight divided by length cubed (i.e. a stylised calculation of density)
bears a closer look. Indeed, when one thinks about it, and bears in mind the
fact that the proportion of bone in a smaller animal is going to be a greater
than in a larger one (this is a surface area compared to volume calculation),
this calculation of the ratio is as perfect as one could wish for a tortoise.
Clearly, it is a much more effective mathematical definition of the graph than
Jackson's "weight divided by length". (This is of course a simplification
of his work anyhow- his definition actually defined that the logarithim of the
weight should be divided by the logarithm of the length)

Practically, of course, out at MOTs (health checks), this has even more significant
effects. Jackson's wt/L calculation is weight specific, not a generalised ratio,
and therefore has little use when examining a specific tortoise.

Let us consider an example:

Two tortoises are presented for weighing prior to hibernation. One is going
to be dangerously underweight, the other, OK.

Tort1 weighs 1300g and measures 180mm

Tort2 weighs 1700g and measures 235mm

Which should not be hibernated:

Using weight divided by length:

Tort1 is calculated to be: 7.2

Tort2 is calculated to be: 7.2

Using weight divided by length cubed:

Tort1 is calculated to be: 0.22

Tort2 is calculated to be: 0.14

The first calculation is of no help. The second quite clear- the lower the value, the more serious the situation. Indeed, if these measurements are plotted on Jackson's graph, one is on the "average" line, one on the "low" line. Only the "Bone" method picks this up without the use of the graph itself.

The value of this method for other species is significant. Without the dammning nature of the graph, with its "above and below" lines, ratios can be observed, and given, without panicking the owner. Each individual tortoise can be tracked over time by its ratio, giving a clear indication of weight gain/loss, regardless of growth. This gives a much clearer impression of an individual tortoise, and factors in such things as gender, body shape, natural obesity etc. Obviously, the exact range of safe ratios for different species are not yet specified in any detail, and are left to the judgement of our MOT teams, but, if this method became BCG "law", then MOTs could swiftly provide more data for standardisation than Blakey, Kirkwood or even Jackson had ever dreamed of!

Paul Coleman