On July 28, 2026, Berkeley’s City Council approved a new Hopkins Street bike-lane configuration on the stated premise that the Fire Department had found it compatible with emergency access. Four hours before the vote, the fire chief told Council that his department had not reviewed that configuration.
This explains one decision, from the beginning, assuming you know nothing about Hopkins Street or Berkeley’s council. Every quotation comes from the public record.
A Class IV separated bikeway — also called a protected bike lane or a cycle track — is a bike lane physically separated from car traffic by a barrier: posts, a planter, a curb, or a row of parked cars.
Its purpose is to provide cyclists greater protection than paint alone. It also takes up more room. On a street of fixed width, that room has to come from somewhere: usually parking spaces, sometimes the width of the driving lanes themselves.
That last part is why the fire department gets involved.
Fire engines are big. To reach a burning building, a truck needs enough clear width to drive down the street, and enough room at the curb to set up and work. If you narrow a street’s driving lanes to make space for a bike lane, you may have made it harder for a fire engine to get through.
Hopkins Street matters more than most. It is a designated Emergency Access and Evacuation Route — one of the roads the city expects to use to move fire equipment in and residents out during a wildfire. Berkeley’s own wildfire plan identifies narrow streets and constrained evacuation capacity as core vulnerabilities of the city.
So on Hopkins, Fire Department review is not a formality. It is the technical check directed at the specific emergency-access and evacuation risks raised by this design.
Berkeley had been planning bike improvements on Hopkins since 2018, and the Council approved a conceptual design in 2022.
Then the City Manager circulated a memo placing the project on indefinite pause. Three problems: lane widths had not been resolved; the fire code official had never approved the traffic-calming devices; and nobody could quantify what the project would do to evacuation times.
The project went into an indefinite pause. Fire department concerns were the reason.
Three years later, Councilmember Rashi Kesarwani brought forward a proposal to finally repave Hopkins — with pedestrian safety improvements only. No protected bike lanes.
Her own later explanation for why: the pedestrian-only version “was based on our prior understanding of the need for significant infrastructure upgrades to meet concerns raised by the Fire Department.”
In plain terms: under the understanding then in place, adding protected bike lanes would require costly infrastructure work to resolve Fire Department concerns. Estimated cost of the pedestrian-only version: about $11 million.
A council committee voted to send her item to the full Council with a negative recommendation unless she added protected bike lanes. She was being told by fellow councilmembers that pedestrian-only was not good enough.
At that same meeting, staff presented a document titled “Hopkins Street Alternatives 1 and 2 Overview.” This document is the hinge of the whole story, so it is worth being exact about what was in it.
| Design | What it was | What fire said |
|---|---|---|
| Alternative 1 | A two-way cycle track — bikes traveling both directions on one side of the street | “Alternative 1 retains workable emergency access overall.” |
| Alternative 2 | Pedestrian safety only — crosswalks, refuge islands, no protected bike lane | “Alternative 2 retains workable emergency access overall.” |
The Fire Department reviewed those two specific alternatives and found that each retained workable emergency access. It also confirmed a general lane-width standard: 11 feet preferred and 10.5 feet minimum.
Kesarwani came back with a new version. Not Alternative 1. Not Alternative 2. A third design:
This is a genuinely different layout from Alternative 1’s two-way track on one side. Different lane positions, different curb impacts, a different distribution of the street’s width.
Here is what her written proposal said, on page 7, under her own office letterhead:
“Updated Review by Fire Department Shows Class IV Separated Bikeways Can Retain Workable Emergency Access.”
Agenda Packet, July 28 2026, Item 27, page 7 of 28
It went on to say the fire safety objections that had paused the project since 2023 “no longer exist.”
And the very first line of her formal recommendation — the operative clause the Council would vote on — began:
“Given new information presented by the Fire Department that pedestrian and bike safety infrastructure on Hopkins Street ‘retains workable emergency access’…”
The fire finding was not background detail. It was the proposal’s stated answer to the objection that had stalled protected bike lanes since 2023, and it appeared in the recommendation’s operative opening clause.
It also changed the cost comparison. The proposal argued that protected bike lanes could now be added “without the need for costly infrastructure upgrades” previously thought necessary to satisfy Fire Department concerns. On that premise, Council was asked to select a design estimated at $14–15 million rather than the roughly $11 million pedestrian-only version — an increase the item justified as worth bearing.
How do you get from what fire approved to what the proposal claimed?
The logic: Fire approved 11-foot and 10.5-foot lane widths → this Class IV design uses those widths → therefore fire’s approval covers this design.
The gap: Lane width is one input to emergency access. It is not the whole question. Where the barriers and bike lanes sit, how apparatus would position and operate, and what the full configuration could mean for emergency access or evacuation flow — none of that follows from a lane-width number.
And critically: the document cited as the source does not say what the heading says. It names Alternative 1 and Alternative 2. It evaluates a two-way cycle track. It never evaluates one-way lanes on both sides, because that design did not exist when it was written.
The meeting ran past midnight. Roughly three hours of public comment. Item 27 was why the room was full.
1 HOUR 31 MINUTES IN — a councilmember asked Fire Chief David Sprague directly for the fire department’s position. His answer:
“FIRE has reviewed and approved the traffic calming features in the Hopkins Alternative 1 and 2… We also generally have confirmed that lane widths… are acceptable.
“However, for the new… proposed configuration that you’re discussing tonight… There is not a design yet, so we have not seen that, we have not reviewed it, and we would need a complete design set from Public Works.
“Once that’s available, we can evaluate… potential impacts to evacuation times, and confirm that design maintains workable access for day-to-day fire and medical emergencies.”
Fire Chief David Sprague, Council transcript, 01:31:48 elapsed
Asked how long a review would take, he said two to four weeks, once Public Works produced a design set.
So the fire chief told the Council, in open session and in response to a direct question, that the proposal’s claim of an updated Fire Department review did not apply to the new configuration before them.
He had also stated the same point in writing. A member of the public read his letter aloud that night: “The newly proposed configuration involving one-way protected bicycle lanes on both sides of Hopkins is a different design, and the fire department has not reviewed that configuration.”
Between the chief’s correction and the vote, four hours passed. In those four hours:
| The proposal was not amended |
| The page 7 heading was not corrected |
| The recommendation’s opening clause — “Given new information presented by the Fire Department” — was left standing |
| Nobody moved to defer the design commitment long enough for the two-to-four-week review the chief described |
A substitute motion was offered that would have continued studying alternatives in parallel rather than committing exclusively to the proposed configuration. It failed 4–5. An earlier substitute, removing the unreviewed alignment, failed 3–6.
6 HOURS 15 MINUTES IN — the Council adopted the plan 7–2.
Nine members had the same four hours and the same information. Their responses distributed like this:
| What a member did after the correction | How many |
|---|---|
| Moved a substitute that would have cured the defect | 2 |
| Backed a cure, then voted for the uncured item when it failed | 1 |
| Voted for the item without seeking a cure | 6 |
| Amended the recommendation, corrected the heading, or sought a continuance | 0 |
The last row is the finding.
Two members did the thing the process theoretically allows: they moved alternatives that would have preserved the comparison or required the analysis first. Both failed. One member voted for a cure and then, ninety seconds after it failed, voted for the item the cure was meant to fix.
Nobody amended the document. That option — strike the heading, qualify the claim, continue the item two weeks for the review the chief said he needed — was available to every member the entire time, cost nothing procedurally, and was never exercised.
The natural reading is that somebody should have caught it. That reading is available, and it is not the useful one, because the process did not require anyone to.
Nothing in Berkeley’s procedure obliges the author of an item to amend a recommendation when its factual predicate is contradicted in open session. Nothing obliges the clerk to flag it. Nothing stops a vote on a document whose cited source has just been shown not to say what the document claims.
The substitute motion is the only available instrument, and a substitute is an all-or-nothing swap: it asks members to replace the whole item rather than repair the defective clause. When a substitute fails, the defect is not addressed — it is ratified, because the only alternative on the floor was rejection of the entire item.
That is the mechanism. A correction entered the record, was heard, was not disputed, and had no procedural consequence, because the system has no step where a contradicted premise must be reconciled before the vote.
A premise contradicted on the record must be reconciled before the vote. Where the department that owns a technical finding states in open session that the finding does not cover the design being approved, the item is incomplete rather than merely contested, and incomplete items are returned.
A recommendation needs a repair instrument. The gap this record exposes is procedural: a council needs a way to amend a defective clause without moving to replace the entire item. Two to four weeks was the stated cost of doing this correctly.
A citation is not a checkpoint. Fire review functioned here as a footnote rather than a gate. A review cited as the basis for a decision should be a review of the thing being decided.
Berkeley’s nine councilmembers cannot personally verify fire code compliance or model evacuation times. They depend on the written record in front of them being an accurate account of what the professionals concluded. Reliable staff and sponsor representations are indispensable to a part-time council governing technically complex matters.
The responsible department head said so aloud roughly four hours before the vote, in answer to a direct question. Council nevertheless committed to advancing the new configuration — estimated at roughly $3–4 million more than the pedestrian-only version — without correcting the written premise or waiting for the Fire Department review the chief said was still required.
A councilmember may argue that a design will satisfy the fire department. A councilmember may not tell the Council that the fire department already has.