No. 12 · February 2026
Investigation · Water
Verrow Water’s treatment works at Marlow Bank has discharged into the River Sennet in dry weather since 2014. The company’s own returns record 3,412 hours of it. We spent fourteen months on the riverbank, in the permit archive and in a laboratory in Bristol, and found a pipe that was never supposed to be open.
Reported over fourteen months
Summary
Four things this investigation establishes, each of them from a document, a return or a laboratory result rather than from an account given to us.
01
Verrow Water’s own event-duration returns record 3,412 hours of discharge from the Marlow Bank storm overflow between January 2019 and December 2024 on days the Met Office logged no measurable rainfall at Ashcombe. That is 142 days of continuous running, in dry weather, from a pipe permitted to open only in wet.
02
Permit EPR/QB3947TT allows discharge “during and immediately following rainfall”. From September 2014 Verrow Water applied that phrase to a rolling 72-hour window after any rain of 2mm or more — an interpretation the regulator confirmed to us in writing it had never agreed, and had never been asked to.
03
Of 214 samples we drew at six points over nine months, 61 exceeded the permit’s ammoniacal nitrogen limit at the point of discharge. The highest, taken at 06:40 on 14 March 2025 after eleven dry days, measured 8.4 mg/l against a consented 0.75 mg/l.
04
Ashcombe Town Council, the two schools that swim at Coldwell Reach and the Sennet Angling Club have never received notice of a discharge. Verrow Water is under no duty to give one. It also holds no record of ever having done so, in eleven years and 4,118 logged events.
Stage one
Step 01
The outfall is a 1.2-metre concrete pipe set into the north bank of the Sennet, four hundred metres upstream of the footbridge at Marlow Bank and nine kilometres above Ashcombe. It is not signed. It appears on no public map of the river, and the nearest thing to a public record of it is a six-digit reference in a spreadsheet the company files once a year.
We first stood in front of it on 2 December 2024, on a dry Monday afternoon, and it was running.
Step 02
Behind the grating is a weir. When the works takes more flow than it can treat, the excess spills over that weir and out through the pipe — a storm overflow, legal, and in a wet enough hour the alternative to sewage backing up into houses in the valley.
Condition 3.2 of the site’s permit allows this “during and immediately following rainfall”. What the condition does not do, and what eleven years of correspondence shows nobody ever pinned down, is say how long “immediately following” lasts.
Step 03
Verrow Water answered that question itself, in an operating note dated 11 September 2014: 72 hours after any rainfall of 2mm or more. Set against Met Office records for the Ashcombe gauge, that window covers most of a British year.
Matching all 4,118 logged events to daily rainfall gives the figure at the top of this piece: 3,412 hours of discharge on days with no measurable rain at all.
Stage two
Step 01
Between June 2024 and March 2025 we drew 214 samples from six points: two above the outfall, one at the pipe, and three below it at 40 metres, at Coldwell Reach and at the weir in Ashcombe. Every bottle was sealed on the bank, logged, and driven the same day to Kestrel Analytical in Bristol, an accredited laboratory that had no other relationship with this story.
Step 02
Coldwell Reach is a gravel bend three kilometres below the outfall where two Ashcombe schools have taught swimming since the 1960s. We sampled it 39 times. On 14 March 2025, after eleven dry days, the water there carried 41,000 colony-forming units of E. coli per 100ml — forty-one times the threshold at which a designated bathing water is closed.
It is not a designated bathing water. Nobody has ever applied for it to be one.
Step 03
At the pipe itself that morning, ammoniacal nitrogen measured 8.4 mg/l. The permit consents 0.75. Sixty-one of our 214 samples exceeded that limit, and 54 of the 61 were drawn on days with no rainfall in the preceding 72 hours.
You can smell the difference from the towpath. You don’t need a laboratory. What you need is somebody who will write it down.
Iris Fenwick, 73, has swum at Coldwell Reach since 1961
Stage three
Step 01
Nine kilometres of river separate the pipe from the town, and the Sennet does a good deal of work in them. By the weir at Ashcombe our ammonia readings had fallen by four fifths. The bacterial counts had not: the median at the weir across nine months was 3,900 cfu/100ml, against 210 above the outfall.
Step 02
Sennet Angling Club has kept a hardbacked fish log since 1974. Its counts of brown trout on the two-mile stretch below Marlow Bank fall from a 1990s average of 61 a season to nine in 2024. Above the outfall, on water the same club fishes, the same count is stable.
We wrote to them eleven times. We were told, each time, that the site was compliant. It was compliant. That was the problem.
Danny Ubido, secretary, Sennet Angling Club
Step 03
Ashcombe is 9,400 people and every street in it runs downhill to the river. Two schools swim at Coldwell Reach in the summer term. The town’s open-water club meets at the weir on Wednesday evenings from May.
None of them were ever told when the pipe was open, because there is no mechanism by which anyone would be. That is not a failure of this company. It is what the system was built to do, and it has done it here for eleven years.
Documents
Obtained under environmental information regulations and from a source inside the company. Names of junior staff are withheld; the redactions below are ours, and they lift as you read.
Exhibit 04
Environmental permit EPR/QB3947TT · condition 3.2 · issued 4 September 2014
The operator may discharge from Outfall MB-0417 to the River Sennet during and immediately following rainfall, provided that the flow through the works is maintained at not less than 3.6 l/s and that each event is recorded in the event duration monitoring return required by condition 7.1.
No definition of “immediately following” appears anywhere in the permit or its schedules.
Exhibit 07
Internal email · Verrow Water operations · 19 August 2022, 17:44
Marlow Bank has been over the weir more or less continuously since 2 August. We are inside the 72-hour rule on paper because of the storm on the 30th, but the honest answer is the inlet pumps have been down since June.
I would rather not put the hours in writing. If anyone asks for the raw telemetry before the year end, send it to me first.
Verrow Water does not dispute the authenticity of this email. Its statement is below.
Exhibit 11
Event duration monitoring return · Marlow Bank · calendar year 2024
| Event | Start | Hours | Rain, 72h |
|---|---|---|---|
| 2024-118 | 14 Jun, 04:10 | 61.4 | 0.0 mm |
| 2024-142 | 2 Jul, 22:55 | 204.9 | 0.0 mm |
| 2024-207 | 19 Aug, 06:20 | 311.2 | 2.4 mm |
| 2024-311 | 27 Nov, 13:05 | 96.7 | 0.0 mm |
Four of the 214 events filed for 2024. The full return, and our matching of it to rainfall, is published alongside this piece.
Right of reply
“Marlow Bank has operated within the terms of its environmental permit throughout the period in question, and every discharge event has been recorded and returned to the regulator as required. The interpretation of condition 3.2 applied at the site was adopted in 2014 on engineering advice and has never been challenged by the regulator. We accept that the inlet pumping capacity at Marlow Bank has been inadequate for the catchment it now serves, and a £41m upgrade will complete in 2028.”
Verrow Water, written statement, 21 January 2026
Method
Everything above can be checked. This is where each part of it came from, and what we could not establish.
All 4,118 event duration monitoring records for Marlow Bank, 2014 to 2024, obtained under the Environmental Information Regulations in three tranches between March and September 2025. We matched every event to the Met Office daily rainfall series for the Ashcombe gauge, and published both datasets and the matching script alongside this piece.
214 samples, six points, June 2024 to March 2025, drawn to a written protocol reviewed by two freshwater ecologists before we started. Bottles were sealed and labelled on the bank, kept at 4°C, and delivered the same day under a chain-of-custody sheet to Kestrel Analytical, a UKAS-accredited laboratory paid at its standard commercial rate.
Eleven internal documents were provided by a source inside the company, who is known to this publication and whose identity is protected. Each was verified independently — against the permit register, against the returns, or by a second person who had seen it — before it was used.
Thirty-one residents, anglers, swimmers, two former Verrow Water engineers and four town councillors were interviewed between January 2025 and December 2025. Everyone quoted here saw their quotations before publication and confirmed them.
We could not establish who signed off the 2014 operating note; the copy we hold is unsigned and the company declined to say. We cannot attribute the fall in trout counts to the discharge alone: agricultural run-off in the same stretch is unquantified, and we say so rather than implying a link the sampling does not carry.