LVII. On a practical thermometric standard