Relay rail inspection is the testing, classification and record-keeping that decides whether second-hand rail can go back into track. Under 49 CFR 213.237(c), plug rail used in tested track must already have been tested for internal flaws, and plug rail installed after March 25, 2014 may not exceed 30 million gross tons (MGT) since that test.
"Relay rail" and "second-hand rail" mean the same thing. FRA's own notice says so: second-hand rail "is sometimes also referred to as relay rail." The regulation itself uses a third term, plug rail, for a length of rail removed from one location and stored for use as replacement rail somewhere else. Plug rail is the case FRA regulates directly, and it is where most relay rail risk sits.
This guide covers the US regulatory chain behind relay rail inspection: the 2001 Nodaway derailment, FRA Safety Advisory 2006-02 and the 2014 rule that replaced it. It then works through the tonnage calculation, what an inspection should check, how ultrasonic testing (UT) handles worn rail, and what the records must show.
Contents: Why it matters · The regulatory chain · The 2014 rule · Worked example · What gets checked · How UT works on used rail · Classification and traceability · Standards and research · Limits · FAQ
Why does relay rail inspection matter before reuse?
Because a used rail's last test result only describes the rail on the day it was tested. Tonnage carried since that test, time in storage and handling all change the picture, and none of it shows from the outside.
FRA explained the timing risk plainly in 2006. Rail in tested main track "can be relatively free of internal defects if the last test occurred shortly before the rail's removal." Rail removed at the end of a testing cycle, or taken from track that was never subject to 213.237 testing, "is more likely to have defects."
The case that started it: Nodaway, Iowa, 2001
On March 17, 2001, Amtrak's California Zephyr derailed near Nodaway, Iowa, on BNSF Class 4 track. FRA's account records 78 people injured, including one fatal injury.
The broken rail at the point of derailment came from a 15-foot, 6-inch plug. BNSF had installed it in February 2001 after a routine test found internal defects in the original continuous welded rail. The plug was visually inspected but received no ultrasonic inspection before or after installation. The National Transportation Safety Board (NTSB) found it had multiple internal defects, and that it failed by fatigue from cracks associated with internal hydrogen (NTSB report RAB-02-01).
The irony is that BNSF tested that subdivision every 30 days, far more often than the rule then required. The process simply assumed a stored rail had passed its last test in its old location. The NTSB then recommended that FRA "require railroads to conduct ultrasonic or other appropriate inspections to ensure that rail used to replace defective segments of existing rail is free from internal defects" (Recommendation R-02-5).
Key takeaway: frequent testing of the track does not cover a rail that arrived after the last test run.
What did FRA Safety Advisory 2006-02 recommend?
FRA Safety Advisory 2006-02 (71 FR 11700, March 8, 2006) set out recommended practices for testing, classifying and reusing second-hand rail. It was guidance, not a regulation. Its first recommendation was direct:
"FRA recommends that railroads retest for internal rail flaws the entire length of any rail that is removed from track and stored for reuse. The railroad should conduct this retest before that rail carries revenue traffic."
FRA also suggested marking retested rail "fully re-tested" or similar. For railroads without out-of-track testing capability, the advisory offered a minimum classification program instead:
| Safety Advisory 2006-02 recommendation | What it said |
|---|---|
| (1) Retest | Retest the entire length of stored rail before revenue traffic, then mark it as retested |
| (2)(a) Classify | Mark rail as reusable only if it came from 213.237-tested track, is going into 213.237-tested track, and has less than 15 MGT since its last valid test |
| (2)(b) Prohibit | Don't reuse rail from untested track, or rail without a classification marking, in 213.237-tested track |
| (2)(c) Mark defects | Use a highly visible permanent mark, such as score lines from an abrasive saw or torch, on the head at each detected defect, in addition to the 213.237 marking |
Note that the 2006 baseline rule for Class 4 track was 40 MGT or once a year. That figure matters for the next step in the story.
What does 49 CFR 213.237 require for plug rail today?
Plug rail used in track that must be tested under 213.237 must have been tested for internal flaws, and it can't pass 30 MGT across its old and new locations before its next test. Plug rail that doesn't comply limits trains to Class 2 speeds until it is tested.
These requirements came from FRA's final rule Track Safety Standards; Improving Rail Integrity (79 FR 4234, January 24, 2014, effective March 25, 2014). The rule text now sits in 49 CFR 213.237(c):
| Paragraph | Requirement |
|---|---|
| 213.237(c)(1) | Any rail used as replacement plug rail in track that must be tested under 213.237 "must have been tested for internal rail flaws" |
| 213.237(c)(2) | The owner must verify that plug rail installed after March 25, 2014 has not accumulated more than 30 MGT in previous and new locations since its last test, before the next required test |
| 213.237(c)(3) | Non-compliant plug rail in use after March 25, 2014: trains must not exceed Class 2 speeds until the rail is tested |
| 213.237(j)(2) | Plug rail "means a length of rail that has been removed from one track location and stored for future use as a replacement rail at another location" |
Why FRA chose 30 MGT instead of retest-before-use
During the rulemaking, the NTSB objected to the advisory's 15 MGT allowance. Under the old 40 MGT interval, rail with 15 MGT on it could be installed just after a test run and carry up to 55 MGT before its next inspection. The NTSB wanted all plug rail tested immediately before reuse.
Track owners told the working group that they either test rail in track before removal and allow some tonnage, or inspect and certify it after removal. FRA settled between the two positions. The preamble states:
"the final rule requires plug rail to be inspected at the same frequency as conventional rail. This requirement therefore supersedes FRA Safety Advisory 2006-02 and codifies current industry practice"
FRA still added that, all else being equal, it "does recommend that the rail be tested prior to installation in track for reuse." It also made clear the verification applies only to plug rail installed after March 25, 2014, not to plugs already absorbed into normal test cycles.
| Approach | Tonnage a reused rail could carry between tests |
|---|---|
| SA 2006-02 classification, old 40 MGT rule | Up to 15 + 40 = 55 MGT (the NTSB's concern) |
| Current 213.237(c)(2) | 30 MGT total, old and new locations combined |
| Retest before installation (FRA's recommendation) | Clock starts at zero on installation; the segment's normal interval then applies |
Key takeaway: the advisory is history. Cite 213.237(c) in procedures and contracts, and treat retest-before-use as FRA's recommended practice, not the legal minimum.
How do you check the 30 MGT limit on plug rail?
Add the tonnage the rail carried in its old location after its last test to the tonnage it will carry in its new location before the next scheduled test. If the total exceeds 30 MGT, the rail needs testing before or soon after installation.
Worked example
A track owner wants to install a 39 ft plug, cut from Segment A during a rail change, into Segment B. Both are Class 4 freight segments.
| Step | Data | Result |
|---|---|---|
| 1. Old location | Segment A carries 36 MGT a year. The plug's last valid test was 4 months before removal | 36 × 4/12 = 12 MGT already accumulated |
| 2. Storage | 6 months in the yard | No added tonnage, but the test is now 10 months old |
| 3. New location | Segment B carries 24 MGT a year and was last tested 25 days before installation | Next test due at 370 days (30 MGT would take 15 months), so about 345 days after installation |
| 4. Tonnage before next test | 24 × 345/365 | ≈ 22.7 MGT |
| 5. Total | 12 + 22.7 | ≈ 34.7 MGT, over the 30 MGT limit |
| 6. Allowance | 30 − 12 = 18 MGT; 18/24 × 365 | The plug must be retested within about 274 days of installation |
The owner has three practical choices:
- Retest the plug before installation. The 12 MGT drops to zero, and Segment B's normal cycle covers it. This is FRA's recommended practice.
- Schedule a test over the plug location within about 274 days and record it against the plug.
- Bring forward Segment B's next test so it falls before the plug reaches 30 MGT.
If none of these happens and the limit passes, 213.237(c)(3) restricts trains over that rail to Class 2 speeds (25 mph freight, 30 mph passenger) until it is tested.
The calculation is only possible if the plug's last test date and its old location's tonnage are known. If they aren't, the only defensible answer is to test the rail before it carries traffic. The interval rules themselves are covered in how often rail must be tested and the 49 CFR 213.237 guide.
What does relay rail inspection check?
A relay rail inspection checks four things: what the rail is, what condition its surface and section are in, what is inside it, and whether it suits the track it is going into. Each needs different evidence.
| Check | Method | Why it matters |
|---|---|---|
| Identity and history | Brand and heat markings, test records, removal location, tonnage data | Needed for the 213.237(c)(2) calculation and for traceability |
| Visual condition | Visual examination of head, web and base | Finds cracks, shelling, burns, corrosion, bends and handling damage |
| Head and gauge-face wear | Rail wear gauge or profile measurement | Decides whether the section suits the new track; heavy wear also affects UT coverage |
| Internal condition | Ultrasonic testing, in track before removal or out of track before reuse | Finds transverse, horizontal and vertical internal defects that visual checks can't |
| Rail ends and bolt holes | Visual and hand UT where joints will be made | Bolt hole cracks and end damage start where the rail will be joined again |
| Fit for intended use | Rail section, weight and profile against the receiving track | A rail suited to a siding may not suit main track |
Visual inspection and internal testing answer different questions. Our guide to ultrasonic vs visual inspection explains the split. A rail can have a clean head and an acceptable profile and still carry a defect such as a detail fracture growing from below the surface.
Wear limits for reuse aren't set in 213.237. They come from the track owner's own standards or the purchase specification, so a relay rail inspection report should state which standard the wear measurements were judged against.
Key takeaway: record the test date and tonnage as carefully as the defects. Without them, a clean UT result can't be applied to 213.237(c).
How does ultrasonic testing work on used rail?
Ultrasonic testing sends high-frequency sound into the rail through a coupled probe and reads the echoes; a reflection where none should be indicates a possible internal discontinuity. On used rail, the main challenge is getting sound in cleanly and reading it correctly.
Probe angles and what they find
Rail test systems use several beam angles because each defect lies in a different plane. A typical configuration is shown below, though probe sets vary by system and manufacturer.
| Defect | Orientation | Typical beam |
|---|---|---|
| Transverse fissure, detail fracture | Crosswise in the head | 70° shear wave, forward and backward |
| Horizontal split head, compound fissure (horizontal part) | Parallel to the running surface | 0° straight beam |
| Vertical split head | Lengthwise, vertical | 0° loss of back-wall echo, plus side-looking or angled probes |
| Bolt hole cracks | Radiating from the hole in the web | Angled probes, often around 37–45° |
A 0° straight beam can't reliably find a transverse defect, and a 70° beam is poorly placed for a flaw parallel to the surface. That is why a single-probe check isn't a full relay rail inspection. More on methods is in internal rail defect detection methods.
In track or out of track?
Testing before removal uses the track owner's normal test run. It gives a valid search under 213.237 and a known date, but tonnage keeps accumulating until the rail comes out. Testing out of track, on the rack or on the ground before reinstallation, needs a hand probe or cart and a way to couple to rail that may be dirty, rusty or uneven. Our guide to hi-rail, cart and walking tests compares the equipment.
Why worn rail is harder to test
Used rail often carries head checks, shelling and spalling, corrosion and scale. Surface damage scatters sound and can mask a defect beneath it. Heavy head wear changes the geometry the probes were set up for.
This isn't only field lore. After the 2006 New Brighton, Pennsylvania derailment, where the NTSB found the probable cause was an undetected detail fracture, FRA's Rail Integrity Task Force was asked to "evaluate the effect of rail head wear, surface conditions and other relevant factors on the acquisition and interpretation of internal rail flaw test results." The NTSB had also recommended methods that "accurately measure rail head wear to ensure that deformation of the head does not affect the accuracy of the measurements" (R-08-11). Both are recorded in the 2014 rule's preamble.
Under 213.237(j), a valid search requires equipment that "performs as intended" and responses "interpreted by a qualified operator." The same discipline should apply out of track. If coupling was lost or the surface masked part of the length, the record should say so, and that length isn't tested.
Preparing used rail for UT
Before testing stored rail:
- Clean loose scale, dirt and debris from the scanning surfaces.
- Calibrate the instrument on a reference standard before testing and confirm it at intervals set by the procedure.
- Use an operator qualified under 49 CFR 213.238 for the specific equipment.
- Test the full length, including rail ends, and record any length that couldn't be searched.
Key takeaway: a clean UT result on used rail is only as good as the coupling and coverage behind it, so the record must show both.
How is relay rail classified and tracked after testing?
After testing, each length is classified as reusable, reusable for limited service, or scrap, and its test status must stay with the rail until installation. SA 2006-02 put traceability at the center of its recommendations, and the 2014 rule made it necessary: 213.237(c)(2) can't be satisfied without knowing when a plug was last tested.
What happens when a defect is found
A detected defect is marked under 213.237(e), which requires "a highly visible marking on both sides of the web and base." For rail still in track, the remedial action comes from the table in 49 CFR 213.113; defect types are explained in our guide to FRA defect codes. For rail out of track, the options are usually:
| Finding | Typical disposition |
|---|---|
| No reportable indication, full length searched | Classify as reusable; mark test status and date |
| Defect near one end | Cut out the defective section, if the owner's procedure allows, then retest the remaining length |
| Defect in the body | Reject from reuse; apply a permanent defect mark so the rail can't drift back into stock |
| Length not fully searched | Hold; retest before classification |
| Wear beyond the receiving track's standard | Classify for lighter service or scrap, per the owner's standard |
What the traceability record should include
| Record item | Why |
|---|---|
| Rail identification (brand, heat, section, length, ID tag) | Links every result to the physical rail |
| Removal location and date | Source of the old-location tonnage |
| Date and type of last internal test, and who performed it | Starting point for the 30 MGT count |
| Tonnage since that test | Input to 213.237(c)(2) |
| Defects found, sizes, and lengths not searched | Shows what the test did and didn't cover |
| Classification and storage location | Prevents mix-ups between reusable and rejected stock |
| Installation location and date | Lets the owner schedule the plug's next test |
Once installed, the plug becomes part of the receiving segment. Its next test is recorded under 49 CFR 213.241 like any other internal rail inspection.
Key takeaway: paint fades and tags fall off in storage, so a permanent defect mark and a matching record matter more than the sensor that found the defect.
Which standards and research apply to relay rail testing?
In the US, the controlling requirement is 49 CFR 213.237(c). Industry standards describe test systems but don't decide reuse.
- ISO 5735-1:2024, Railway infrastructure — Non-destructive testing on rails in track — Part 1: Requirements for ultrasonic testing and evaluation principles, was published on March 14, 2024 (BSI Knowledge listing). National catalog summaries describe it as covering continuous testing of flat-bottom rails of 43 kg/m and above in track, by test vehicle or manually propelled device. The same summaries say it gives no guidance on managing test results and excludes production-plant testing. It isn't referenced in Part 213, so it applies only where a contract or owner procedure calls for it.
- ISO 5948:2018 covers ultrasonic acceptance testing of railway rolling stock material such as wheels, tyres and axles. It isn't a standard for rail in track.
- FRA research on inspection reliability. FRA report DOT/FRA/ORD-26/07, Application of MAPOD to Handheld Ultrasonic Rail Flaw Inspection (ENSCO, July 2026), studied model-assisted probability of detection (MAPOD) for handheld rail UT, the kind of equipment used for out-of-track and spot testing. Its abstract points to further work on the probability of correctly sizing flaws and on adding physical and human factors to the simulations. It is research, not a requirement, and it doesn't publish a detection rate a buyer should quote as a guarantee.
A Canadian case is a reminder that the joint matters as much as the rail. The Transportation Safety Board of Canada concluded in investigation R20W0025 (Guernsey, Saskatchewan, February 2020) that a plug rail joint installed about a week earlier most likely failed under a crude oil train. US readers should treat Canadian rules as context only; Part 213 governs here.
What relay rail inspection cannot tell you
- It can't prove the rail is defect-free. UT finds reflectors above a detection threshold in the orientations the probes cover. Small, tight or badly oriented flaws can be missed, and no regulation requires 100% detection.
- It can't see through masking. Shelling, head checks, corrosion and scale can hide internal defects from the beam. A length that couldn't be searched should be recorded as untested.
- It says nothing about future tonnage. A clean test starts the clock; it doesn't stop it. The 30 MGT limit still applies.
- It doesn't judge the joint. Joint bar condition, bolt torque, tie support and rail end batter after installation are track maintenance questions under other Part 213 sections.
- It doesn't apply everywhere. 213.237 covers Class 3 to 5 track (with conditions on Class 3), and 49 CFR 213.3 excludes plant railroads, isolated tourist lines and rapid transit not connected to the general system. Rail transit is also subject to state safety oversight; in California, the CPUC. On track outside 213.237, the owner's own standard governs relay rail reuse.
- It doesn't replace the welding or joining inspection. A plug installed with thermite welds needs its own weld inspection.
How EMA Industries applies this
EMA Industries performs ultrasonic rail flaw testing for transit agencies, short lines and industrial track owners. We are based in Southern California and mobilize nationwide. Whoever tests your relay rail, the report should tie each length to an ID, a test date and any untested length, so you can show compliance with 213.237(c)(2) before installation. To plan a stockpile test or a plug rail procedure, see our ultrasonic rail testing services or contact EMA Industries.
Related reading: 49 CFR 213.237 rail flaw detection requirements · How often must rail be tested? · FRA defect codes explained · Internal rail defect detection methods
Frequently asked questions
Does relay rail have to be ultrasonically tested before it is installed?
Not always immediately before. Under 49 CFR 213.237(c)(1), plug rail used in track that must be tested under 213.237 must have been tested for internal flaws, either in its old location or after removal. It must also stay under 30 MGT since that test until its next scheduled test. FRA recommends testing before installation when practical.
Is FRA Safety Advisory 2006-02 still in effect?
No. FRA stated in the 2014 Improving Rail Integrity final rule that the plug rail requirement "supersedes FRA Safety Advisory 2006-02." The advisory is still useful background on classification and permanent defect marking, but procedures should cite 213.237(c).
What happens if untested or over-tonnage plug rail is already in track?
Under 213.237(c)(3), trains over that rail must not exceed Class 2 speeds (25 mph freight, 30 mph passenger) until the rail is tested. The fastest fix is usually a hand test over the plug location.
What is the difference between relay rail and plug rail?
Relay rail is the general industry term for second-hand rail. Plug rail is the regulatory term in 213.237(j)(2): rail removed from one location and stored for use as replacement rail at another. Most relay rail used as short replacement lengths is plug rail under the rule.
Can rail from an industrial spur be reused on main track?
FRA's 2006 advisory recommended against reusing rail from track not subject to 213.237 testing in tested track, because its history is unknown. Under the current rule, such rail must be tested for internal flaws before use in tested track, and its wear and section must suit the receiving track.
Sources
- FRA Notice of Safety Advisory 2006-02, 71 FR 11700–11701, March 8, 2006, govinfo
- Track Safety Standards; Improving Rail Integrity, final rule, 79 FR 4234, January 24, 2014, govinfo (Nodaway and New Brighton background; plug rail discussion)
- 49 CFR 213.237, Inspection of rail, eCFR (read on Cornell LII, October 2026)
- 49 CFR 213.238, Qualified operator, eCFR
- 49 CFR 213.241, Inspection records, eCFR
- 49 CFR 213.113, Defective rails, eCFR
- 49 CFR 213.3, Application, eCFR
- NTSB Railroad Accident Brief RAB-02-01 (Nodaway, Iowa, adopted March 5, 2002) and Safety Recommendation R-02-5, as cited in the two Federal Register documents above
- Bruzek, Einbinder and Tunna, Application of MAPOD to Handheld Ultrasonic Rail Flaw Inspection, DOT/FRA/ORD-26/07, FRA, July 2026, ROSA P
- ISO 5735-1:2024 catalog listing, BSI Knowledge
- TSB rail investigation R20W0025, Transportation Safety Board of Canada
