B & D Clays & Chemicals Ltd v Castle Water Ltd: meter readings as conclusive evidence of water supply

The Court of Appeal has dismissed the appeal in B & D Clays & Chemicals Limited v Castle Water Limited [2026] EWCA Civ 171, upholding the county court appeal judge's finding that water meter readings constitute binding evidence of the volume of water supplied where the meter has passed the prescribed statutory accuracy test.

The dispute arose from charges of £28,014.37 levied by Castle Water for water supplied to commercial premises in Mitcham between April 2020 and March 2021. Clays contended that one of the two meters on site was defective, pointing to a significant spike in recorded daily consumption — from 3.5 to 14.2 cubic metres — between July 2018 and July 2020. Both meters were subsequently tested and found to operate within the prescribed limits of error under the Measuring Equipment (Cold-water Meters) Regulations 1988.

The core question before the Court was whether regulation 8 of the Water (Meters) Regulations 1988 creates a comprehensive statutory code — such that a meter reading certified in accordance with the Regulations is conclusive unless the meter fails the prescribed test — or whether a customer may adduce other evidence to challenge a reading from a meter that has passed that test.

The majority, comprising Lewis LJ and Peter Jackson LJ, held that regulation 8 does establish a comprehensive regime. Meter readings proved by certificate under regulation 8(2) constitute evidence of the volume of water supplied unless the meter is shown to register incorrectly, which regulation 8(3) defines by reference to the prescribed testing regime in regulation 6. No other evidence may be adduced to displace those readings or demonstrate that the meter was otherwise defective. Peter Jackson LJ observed that the absence of any express provision for "transient, self-correcting faults" was not mere legislative silence but a deliberate and coherent omission consistent with the scheme's purpose: to correlate charging with consumption using a reliable, uniform mechanism. He also noted the acute disproportionality in this case, where costs across all proceedings approached £600,000 against invoices of less than £30,000.

Snowden LJ reached the same outcome but by a different route on the interpretive question. In his view, regulation 8(1) does not use the language of "conclusive evidence" and regulation 8(3) is not exhaustive as to the means by which a meter may be shown to register incorrectly. He would therefore have allowed the appeal on ground 1, permitting other evidence to be adduced in principle, particularly where an intermittent fault — such as scale build-up causing the dials to advance by two digits rather than one — could produce an aberrant reading that later self-corrects before testing.

However, Snowden LJ agreed with the majority on ground 2: the evidence adduced by Clays was insufficient in any event. The joint expert had not dismantled or internally inspected the meter. His conclusion that "something was not right" rested substantially on information provided by a witness who failed to attend trial and whose evidence therefore carried little weight. An unexplained spike in consumption, combined with speculation about possible mechanical causes, could not on the balance of probabilities establish that the meter had been registering inaccurately.

The judgement confirms — at least by majority — that the statutory testing regime under the 1988 Regulations is the exclusive mechanism for challenging water meter readings, and that the regime's durability over nearly four decades reflects a deliberate policy choice favouring certainty and proportionality over bespoke evidential disputes.

Copyright & permissions