The conventional allometric method entails fitting a straight line to logarithmic transformations of the original bivariate data and then back‐transforming the resulting equation (at least implicitly) to form a two‐parameter power function, Y = a × X^b, on the arithmetic scale. Although the protocol is widely used in contemporary research, it commonly performs poorly and thereby leads investigators to form inaccurate impressions of the dominant pattern in their data. Here I re‐examine the metabolic allometry for six species of carabid beetle to illustrate the problem that arises when pattern in the original data can be described by two (or more) statistically equivalent equations with different functional form. Whereas conventional analyses of data for the beetles yielded only a single descriptive model for each dataset (i.e., the two‐parameter power equation), the more versatile protocol used here fitted two to four statistically equivalent equations (including the two‐parameter power function) to the same sets of observations. Conclusions based on just the power equation estimated by conventional allometry would be misleading because the equation does not afford a unique description for pattern in the data.

中文翻译：

---

关于甲虫的静态代谢异构体的新观点。

传统的变通方法需要将一条直线拟合原始二元数据的对数转换，然后（至少是隐式地）对所得方程进行反变换以形成一个两参数幂函数，Y =  a  ×X ^b，在算术规模上。尽管该协议在当代研究中被广泛使用，但它通常表现不佳，从而导致调查人员在其数据中形成对主导模式的不准确印象。在这里，我重新检查六种甲壳虫的代谢变构，以说明当原始数据中的模式可以由两个（或多个）具有不同功能形式的统计学等价方程描述时出现的问题。常规的甲虫数据分析对于每个数据集仅产生一个描述性模型（即，两参数幂方程），但此处使用的通用性更高的协议拟合了两到四个统计等效方程（包括两参数幂函数）对相同的观察结果。