This is the first part of a three-part paper describing the Railway LED Lantern Test (RLLT). In this first part, the historical context of color vision testing on the railways is reviewed, the construction of the RLLT is described in the context of the recommendations1 of the Commission Internationale de l’Éclairage (CIE) regarding a practical test, and performance data are reported for color vision–normal (CVN) and color vision–deficient (CVD) subjects. In part 2,2 the RLLT, carried out with a 6-m working distance, and the CN Lantern, used on the Canadian Railways, (the only commercially available lanterns based in railway practice) will be compared and their appropriateness as lanterns for color vision standard 1 of the CIE is considered. In part 3,3 the RLLT, carried out with a 3-m working distance, will be compared with the Farnsworth Lantern (FaLant) and the OPTEC 900, which are recommended as appropriate lanterns for CIE Standard 2.
Color vision standards for railway locomotive drivers were almost universal within a few years of Holmgren writing that he “supposed that color-blindness was one of the principal causes of the accident” near Lagerlunda, Sweden in 1875.4 Several authors are listed by Mollon and Cavonius5 as having cast much doubt on this supposition. Mollon and Cavonius carried out a comprehensive review of the evidence and concluded that “the hypothesis remains plausible” and “the signals would have been readily confused by daltonians” but that “there is no firm evidence that color deficiency did cause the collision.” They identified a number of other factors that undoubtedly contributed to the accident. Rightly or wrongly, Holmgren’s supposition led, after other accidents around the world, to the widespread setting of color vision standards in the rail industry.6 Concern about the occupational consequences of “color-blindness” precedes Holmgren. George Wilson wrote extensively in 1855 on the risks of color vision deficiencies and occupations, most notably in railway and marine applications,7 and his work was quoted by James Clerk Maxwell earlier in the same year.8
The form of these color vision tests and standards has varied over the years. The railways used Holmgren’s wool test initially and then a simplified version commissioned from Thomson. The Holmgren-Thomson test is the one typically illustrated today.9 In 1880, Thomson also produced a further modified version for the Pennsylvania Railway Co that was intended to be quicker to perform.9 The front cover of volume 93(1), 2009 of the British Journal of Ophthalmology carries a color image of this test, known as the “Thomson Stick.” The Edridge-Green lantern (1890) is often quoted as being the first lantern test,10,11 and it was certainly intended by Edridge-Green for railway as well as maritime use. There is no evidence that the signal colors used at the time on the railways or on the water were the same or that the lantern colors accurately reflected either practice. In the modern systems, the two fields adopt different color codes. Maritime authorities12 (like aviation13) use a red, green, and white code. Railways generally use a red, yellow, and green code for fixed signaling but a red, green, and white code for handheld signals and torches. For maritime and aviation applications, there are internationally agreed specifications.12,13 However, there appears to be no internationally agreed specifications on railway signal lights. Sometime before 1897, the Donders lantern was in use on the Dutch railways14,15 and Nuel (cited by Williams15) was using a lantern-type test in Belgium. Williams himself introduced his lantern in 1897 and, subsequently, published papers on improved versions.16,17 Cole and Vingrys18 list many of the lantern tests existing up to the time of their study (1982). They list British Rail as a user of the Edridge-Green lantern and the US Railway companies as users of the Williams and Thomson lanterns. As with the Edridge-Green lantern, there is no evidence that the colors used in the Williams lantern had been based on signaling practices and visual examination of available models of the Edridge-Green and Williams lanterns in the University of New South Wales (UNSW) Colour Vision Clinic certainly shows that the colors are not related to any modern railway practice.
Color vision testing was introduced in New South Wales, Australia (NSW) railways in 1886 using a test consisting of colored wools,19 possibly Holmgren’s or Thomson’s wool test or something similar. In 1888, the first Chief Commissioner of NSW Railways (EMG Eddy) introduced a colored card test, which he brought with him from his previous employers, the London and North Western Railway. In 1902, NSW Railways adopted a test using colored (red, green, purple, and yellow) filters in a procedure introduced by the, then, Railway Medical Officer, GPM Woodward. In 1904, the Interstate Conference of Railway Commissioners introduced the Williams lantern and reintroduced the Holmgren Wool test and these two were used until 1915. Later, the NSW Railways changed to the modified Williams lantern (an electrified version of the kerosene-burning lantern) and that was in use in 195020 and as recently as 1976.21 In the latter paper, the author reports the use of the Edridge-Green lantern10 and the “Australian Railways lantern” in other states. Interestingly, there is correspondence in the British Medical Journal in 1910 between Edridge-Green and Taylor.22–24 Edridge-Green was clearly annoyed that Taylor was using the Williams lantern on the NSW Railways and was advocating it as easier to use than the Edridge-Green lantern, which was in use at the time on the Victorian25 and Queensland Railways.
Many other lanterns have been designed from time to time but most were intended for other than railway applications (usually maritime or aviation). The Martin lantern was proposed for transport applications generally.26,27
In 2004, the National Transport Commission of Australia published the National Standard for Health Assessment of Rail Safety Workers.28,29 This document nominated the FaLant for trackside workers but required normal color vision (pass the Ishihara test) for drivers. The FaLant was developed at the Naval Submarine Medical Research Laboratory by Farnsworth.30 It was validated for the color recognition tasks in naval vessels31,32 and has been the subject of ongoing review and remains the test of choice in the US Navy. The FaLant became the test of choice in Australian civil aviation33 and military.21 Several are held in a number of locations around Australia. The FaLant is no longer made but the OPTEC 900 by Stereo Optical of Chicago is accepted as the modern replacement.34,35
The Holmes-Wright Type A lantern36 (intended for aviation use) has been evaluated for railway use37 but found to have an unacceptable level of false negatives compared with a simulated railways task. The Type B lantern (intended for maritime use) has a lower light intensity and hence will show a higher failure rate.
The Canadian National Railways38 lantern is based on Canadian signaling practice with incandescent signals presented as a triplet. Many jurisdictions are now moving to light-emitting diode (LED) signals.
The Fletcher-Evans CAM lantern was produced to replace the Holmes-Wright lanterns (red, green, and white code).39 It also includes a yellow code (ostensibly for “clinical” testing) and there are claims of its suitability for railways but the three operating modes are Holmes-Wright A, Holmes-Wright B, and Clinical. In the Clinical mode, the yellow is additional to the white rather than replacing it. Validation has not extended beyond the four cases reported.39 Its applicability to the railways remains untested.
One consequence of Holmgren’s observation and the subsequent imposition of visual standards is that locomotive drivers became a selected population with, in principle, only normal color vision (unless there is a deliberate attempt to deceive or a poor testing regime). The main exception to this is where an acquired color vision deficiency develops and this has been implicated in accidents.14,40 Comparisons of the safety records of locomotive drivers, with and without a color vision deficiency, are not, as a consequence, possible. It is possible, however, to conclude that, even if normal color vision is assumed for locomotive drivers, there are examples of failures to detect and recognize signals. These events are termed “signal passed at danger” or SPAD. Modern safety practice has reduced the reliance on visual detection and recognition of signals but SPADs remain an identifiable event in safety statistics.41–43 Despite the setting of color vision standards, SPADs have been implicated as contributors in several of the major rail accidents in the United Kingdom since 1994.44,45 Given the existence of issues with color detection and recognition even with (probably) normal color vision, CVD subjects are an added complication. Vingrys and Cole46 conclude, “Clearly, there is sufficient evidence to warrant the retention of color vision standards in those transport industries where the highest standards of safety are expected.”
Because there are no internationally agreed signaling practices, it is not surprising that there are no internationally agreed vision standards for railway workers. The CIE1 has a set of recommendations for color vision standards for transport, including railways. These recommendations acknowledge that it is inappropriate to exclude all people with a CVD from critical positions with color contingent decisions because there are forms of CVD that are sufficiently mild as not to impair the ability to detect and recognize color signals accurately and rapidly. The CIE recommendations47 include three standards.
Standard 1 (Normal Color Vision)
This standard requires normal color vision. It is the appropriate standard when:
- colored signal lights must be recognized at long distances or under adverse visibility conditions;
- surface color codes with more than three colors are used as a primary and important means of conveying information on computer screens or other visual information displays; and
- a failure to see or to identify a colored signal or color code is likely to cause an operational error or accident; high social, economic, or environmental costs are likely to be associated with accidents; or there is a community expectation of the highest standards of safety.
Standard 2 (Defective Color Vision A)
This standard passes persons who have a mild color vision deficiency but have a demonstrated ability to identify signal colors correctly. It fails those persons with a protan defect of color vision who have a much reduced ability to see red signal lights. It is the appropriate standard when:
- recognition of colored signal lights and other color codes is important to safe operation, or red signal lights have to be seen; and
- significant social, economic, or environmental costs may be associated with accidents, or there is a community expectation of high standards of safety.
Standard 2 (Defective Color Vision B)
This standard passes persons who have a color vision deficiency but have a demonstrated ability to recognize color codes at short distances. It is the appropriate standard when:
- surface color codes, including color codes on computer screen displays and large signal lights on control panels, have to be recognized at short distances under good conditions of visibility; and
- economic or environmental costs may be associated with operational errors or accidents.
The CIE recommends that CIE Color Vision Standard 1 (Normal Color Vision) be applied to locomotive drivers, engineers, and signalmen and that CIE Color Vision Standard 2 (Defective Color Vision A) be applied to any other person employed in rail transport such as shunters, trackside, and rail yard workers whose occupation requires the recognition of signals.
The acceptance criteria for each of these standards may be summarized as follows:
- Standard 1 (normal color vision): A pass on the Ishihara test plus another pseudoisochromatic plate test selected from a table of seven or results of anomaloscopy or a pass on a lantern test with a high degree of difficulty.
- Standard 2 (defective color vision A): A pass on a lantern test (examples are given) and protans are excluded.
- Standard 3 (defective color vision B): A pass on the Farnsworth D-15 test or compliance with an anomaloscope-based requirement.
The present study was carried out against a background of the, then, current Australian medical standards.28,29 These required, in summary, normal color vision (pass on the Ishihara test) for locomotive drivers and a pass on the FaLant or OPTEC 900 lanterns for other trackside workers. The regulations did allow, in the case of failure of a lantern test, for a practical test. The CIE1 describes practical tests as “superficially attractive” but caution that “a practical test must be devised, validated and administered with the utmost care and caution” and observe that “it is very difficult to devise a practical test that has demonstrated reliability and validity.” They, further, set down seven principles for the design and conduct of practical tests. There are a number of consequences of these principles that include, most notably, that they will carry a high economic cost and that results obtained at one site may not be applicable to another site, such that further practical tests may be needed if the worker changes work location.
The Railway LED Lantern Test
The RLLT was constructed to represent the modern railway-signaling practice using LED signals.
The prototype test was designed using the seven guiding principles of the CIE1:
- The colors to be identified should conform to the color specification for the color code system used in practice but should be chosen to be those colors, within the specification, most likely to be confused by persons with defective color vision.
- The colors should be identified from the longest distance from which color recognition is expected in practice.
- The intensity of colored signal lights or the luminance of colored surface codes should be close to the lowest likely to occur in practice.
- Care must be taken to ensure that there are no noncolor cues to the identification of the color unless these noncolor cues are always present in practice.
- Repeated trials must be given.
- The practical test must be given with formally defined instructions and set procedures followed.
- The practical test must be validated.
Aldridge Railway Signals Pty Ltd provided the LEDs from their stock of LEDs used in signal construction. In this way, the accuracy of the representation of the color of signals in use was assured, which satisfies principle 1. The range of luminous intensities was set on the basis of measurements made on every model of LED signal in stock with RailCorp (now known as Sydney Trains) before installation. When viewed at 6 m, the intensities represent the task of a driver to detect and recognize signals at the reported necessary stopping distance of a high-speed country train (1.6 km). This represents the worst-case scenario plus some variation to be expected from signal to signal. Because the signal is trackside and track curves will be modest (or taken at slower speed), the on-axis luminous intensity will be the important measure. Undertaking the test at 3 m represents the shorter viewing distances but also reflects the need to view signals more obliquely (with the consequential variation of luminous intensity). The 3-m distance is representative of the longest platform in use in NSW (Albury, 500 m) and should be applicable to the longer platforms in Australia (Flinders St, Melbourne, Victoria, 708 m; Kalgoorlie, Western Australia, 760 m; and East Perth, Western Australia, 770 m). In other jurisdictions, any significantly different distance in the workplace can be represented by an appropriately chosen test distance for the lantern. The signals are presented in vertical pairs (which represents the suburban signaling practice of showing the next two sections of track) and some are presented singly. The signals are presented against a black background.
The tests reported here were carried out using a prototype RLLT. The commercially available lantern reproduces the same chromaticities and, with two exceptions detailed later, the same luminous intensities as the first prototype. All lights were presented for 2.0 ± 0.1 seconds. In the prototypes, the presentation of the first 12 pairs of lights and the second 12 pairs of lights were on opposite faces of the lantern, and on the commercially available lantern, they are all presented on one face.
The chromaticities and luminous intensities are considered to be commercially sensitive and not detailed here.
The prototype used for testing has been illustrated before48; the production model is shown in Fig. 1.
Subjects were chosen to meet the following criteria:
- Subjects were between the ages of 18 and 60 years.
- Their corrected visual acuity was no worse than 6/9 binocularly.
- The subjects had healthy eyes with no current and/or active ocular pathology.
- The subject had to be within 1 hour drive from UNSW.
- They were recruited using the patient database of the Color Vision Clinic in the School of Optometry, UNSW, and via advertisements on campus.
- Their correcting spectacles or contact lenses did not have any visible tint.
- Subjects had to be able to speak sufficient English to undertake the test.
- Both men and women were recruited; however, the participants in all sample groups were predominantly male and this was also the RailCorp employment demographic at the time.
The subjects were either classified as “Experienced” or “Naïve” in observing rail signals.
- Naïve: They did not have any significant experience with identifying colored signals in commercial work. The subjects recruited had either normal color vision (n = 16; 7 male and 9 female subjects) or a color vision deficiency (n = 37; 34 male and 3 female subjects).
- Experienced: They did have significant experience with identifying colored signals in a commercial work requiring normal color vision (n = 30; 26 male and 4 female subjects; the preponderance of males reflects the demographic of RailCorp workers).
Subjects were classified as “Normal” or “CVD.”
- Normal (CIE Standard 1): Passes both the Ishihara and the anomaloscope (n = 16 naïve and n = 30 experienced).
- CVD: Examined in UNSW color vision clinic (n = 37; 34 male and 3 female subjects).
The CVD subjects were then classified as “anomalous trichromat” (n = 31; 28 male and 3 female subjects) or “dichromat” (n = 6; all male) on the basis of the anomaloscope.
The anomalous trichromat subjects were further divided into three categories:
- Mild: pass both the Farnsworth-Munsell D-15 and the FaLant (CIE CVD Standard 2) (n = 9; 8 male and 1 female subject).
- Moderate: pass the D-15 but fail the FaLant (CIE CVD Standard 3) (n = 11; 9 male and 2 female subjects).
- Strong: fail both the D-15 and the FaLant (n = 11; all male).
The general color vision examination (including the Ishihara test and Neitz anomaloscope for all subjects and the Farnsworth-Munsell D-15 and FaLant for CVD subjects) was carried out using the routine procedures of the color vision clinic.
For the RLLT, the test was interleaved with other lantern tests that are not reported here.
The subjects were seated in a normally lit room (≈300 lux) without significant glare and including no visible windows. They were seated at 6 m from the RLLT and viewed the RLLT binocularly. They were instructed:
“You will be shown a pair of lights, one above the other for 2 seconds. I want you to tell me what colors you see. Tell me the top one first. The only colors you will be shown are red, yellow and green and these are the only color names that you may use. In some cases there is only one light, in which case respond ‘no light’ again in the order of top first and then bottom.”
The pairs of colors were displayed once only for 2 seconds and the response was recorded. All 24 presentations were shown in ascending numerical order. The subject was then moved to 3 m from the RLLT and the process was repeated in reverse order.
Misnaming a color, failing to see a color, or naming a blank as a color were all recorded as errors.
The number of subjects with various errors in misnaming colors, missed colors, and blanks named as colors is set out in Fig. 2 for both the 6-m and 3-m testing distances. Breaking that down by the type of error, Table 1 shows the errors at 6 m testing distance classified by type of error. The most common error by CVN subjects is misnaming red as yellow (2.3% [confidence interval (CI), 1.3 to 3.7%] of reds). This is not an unknown phenomenon with small signal lights and color normal observers.49,50 On the other hand, some tinted lenses may cause yellow to look red.51 Misnaming of green as yellow (0.3% [CI, 0.04 to 1.1%] of greens) was the only other CVN subject color-based error. Color vision–normal subjects failed to see 4.3% of reds, although this is addressed later in experiment 2.
The pattern of misnamings by CVD subjects is quite different. The most common error is misnaming yellow (51.0% [CI, 46.5 to 55.4%]) as either red (31.3% [CI, 27.3 to 35.6%]) or green (19.6% [CI, 16.3 to 23.4%]). Green was misnamed the next most frequently (24.8% [CI, 21.1 to 28.8%]), as red (13.5% [CI, 10.6 to 16.8%]) and yellow (11.1% [CI, 8.5 to 14.2%]). Red errors were the least common misnaming (18.3% [CI, 15.0 to 21.9%]) and named as green (12.5% [CI, 9.7 to 15.7%]) or, less often, as yellow (5.8% [CI, 3.9 to 8.2%]). The color missed most often, by far, was red (22.8% [CI, 19.2 to 26.7%]).
Table 2 shows the errors at 3 m testing distance classified by type of error. In all cases, the frequency of error is slightly changed for CVN subjects. This may mean that the errors arise from inattention, which will not differ between testing distance, rather than being color vision–based errors. The pattern of error by the CVD subjects at 3 m is similar to that at 6 m, although the absolute frequencies are reduced to between 30 and 70% of the 6-m results. The tendency for CVN subjects to miss red signals is still evident and this is addressed later.
The relationship between the average number of errors at a 6-m testing distance and the degree of color vision deficiency is shown in Fig. 3. The relationship between the average number of errors at a 3-m testing distance and the degree of color vision deficiency is shown in Fig. 4.
It was noted that two of the red lights were not seen particularly frequently by both CVN and CVD subjects but were never the sole mistake of CVD subjects (i.e., they also misnamed colors or named blanks). These two lights were the reds with the two lowest luminous intensities. In an attempt to reduce the false positives that these two lights gave rise to, the same experiment was run with the luminous intensity of those two lights doubled.
A second prototype RLLT was constructed with 10-turn rotary potentiometers instead of the fixed resistors of the first prototype. The potentiometers were adjusted such that the luminous intensities of the second prototype matched those of the first prototype (within ±10%) with the exception of the two lowest intensity red lights, which were set with doubled luminous intensities.
Because the CVD subjects in experiment 1 made other mistakes that would have resulted in their failing the test, only CVN subjects were recruited for this part of the study. They were all naïve subjects and the inclusion criteria are the same as for the first part of the study. A total of 106 naïve CVN subjects (77 male and 29 female subjects) were recruited with an age range of 16 to 60 years.
The same procedure as in experiment 1 was adopted.
The 106 subjects had a mean (±SD) age of 27.8 (±6.2) years.
At 6 m, one subject only missed one red signal (0.1% of presentations, CI <0.4%). There were no misnamings of lights or of blank presentations. At 3 m, all subjects performed the test without any error. These results are annotated with an asterisk (*) in Tables 1 and 2.
On the basis of the performance of the CVN subjects on the RLLT, a pass criterion of less than or equal to two misnamings, less than or equal to one green or yellow missed, and less than or equal to one extra was adopted, based on covering at least the 95th percentile normal. No missed or misnamed reds were permitted given the greater safety implications of missing a red signal.
The change of intensity of the two lights removed the problem of normal subjects failing to see some red lights.
Fig. 2 shows the observed difference in performance between CVN and most CVD subjects, which was as expected. Performance of the test at 6 m is more difficult than at 3 m. At 6 m, all the CVD subjects fail. In the modified prototype, one CVN subject fails because of missing one red light. At 3 m, many of the mild CVD subjects can pass the test and demonstrate that they are able to detect and recognize the lights at the shorter distances representative of tasks other than locomotive driving. Figs. 3 and 4 show the expected relationship with the extent of the color vision deficiency. No CVD subjects classed as moderate or greater passed at either 6 m or 3 m testing distance; thus, the test has effectively excluded these persons from these tasks.
The results demonstrate that the RLLT can be used to identify those workers who have the ability to recognize signal lights sufficiently accurately as to be considered acceptable in the categories.
The RLLT embodies the colors of modern LED signaling practice and luminous intensities based on practical situations such that it has relevance to and consistency with rail signaling practice and the task of a locomotive driver. At a 6-m testing distance, it may be used as “a lantern test with a high degree of difficulty” in the CIE42 schema and acts a simulated practical test that may be administered under controlled conditions in a way that satisfies the seven CIE guiding principles.1 Similarly, at a 3-m testing distance, where it simulates the task of other trackside workers, it may be used for the assessment of those workers in a practicable way with a demonstrated validity.
In October 2012, the Australian National Transport Commission published a second edition of the National Standard for Health Assessment of Rail Safety Workers.52 This contains substantial detail on risk assessment whereby the task is determined to require normal color vision, color defective Safe A vision, or color defective Safe B vision or to conclude that color vision is not required. The document then sets the requirements as:
- Normal color vision required: Pass the Ishihara test (<3 errors in 12 screening plates) or pass the RLLT at 6 m.
- Color defective safe A vision required: Pass the Ishihara test (as above) or pass RLLT at 3 m or pass FaLant.
- Color defective safe B vision required: Pass the Ishihara test (as above) or pass the Farnsworth D-15 test.
This clearly follows the CIE1 schema using the RLLT at 6 m testing distance as the “lantern test with a high degree of difficulty” and, along with the FaLant, at 3 m as an appropriate lantern. The RLLT is a simulation of modern railway signals and there is now no option provided for a separate practical color vision test. Under the transitional arrangements (clause 26.646), workers who had previously been accepted on the basis of the practical test are permitted to continue in their position but will have to meet the current standard should they apply for a position with different color vision demands.
The RLLT is based on NSW signaling practice, but with the increasing adoption of LED signals, there will be much less variation in signal colors and will have similar validity for other jurisdictions. This would be assisted if railway-signaling practice uses the colors recommended by the CIE.53 Intensity is, perhaps, less likely to be highly standardized but variations in different jurisdictions can be accommodated by the setting of different test distances to reflect the different luminous intensity requirements in different jurisdictions and after an assessment of the prevailing practices.
The RLLT has been shown to have validity and embodies a direct link to railway-signaling practice. It has been successfully incorporated into a railway health standard.47 In the same way, it fits into the CIE recommendations for color vision standards on the railways.42
This study was supported, in part, by a grant from RailCorp NSW (now Sydney Trains). The assistance of Clair Taylor in data collection in experiment 1 is appreciated.
The RLLT is available from ART Electronics (www.rllt.com.au). Sydney Trains own the rights to the RLLT.
Received January 31, 2014; accepted August 19, 2014.
1. Commission Internationale de l’Éclairage. International Recommendations for Colour Vision Requirements for Transport
. Vienna, Austria: Commission Internationale de l’Éclairage; 2001.
2. Dain SJ, Casolin A, Long J. Color vision
and the railways: Part 2. Comparison of the CN Lantern used on the Canadian Railways and Railway LED lantern tests
. Optom Vis Sci 2015; 92: 147–51.
3. Dain SJ, Casolin A, Long J. Color vision
and the railways: Part 3. Comparison of FaLant, OPTEC 900 and Railway LED lantern tests
. Optom Vis Sci 2015; 92: 152–6.
4. Holmgren F. Color-blindness in its relation to accidents by rail and sea. In: Board of Regents of the Smithsonian Institution. Annual Report of the Board of Regents of the Smithsonian Institution. Washington, D.C.: Government Printing Office; 1878: 131–95.
5. Mollon JD, Cavonius LR. The Lagerlunda collision and the introduction of color vision
testing. Surv Ophthalmol 2012; 57: 178–94.
6. Burnham JC. Accident Prone: A History of Technology, Psychology, and Misfits of the Machine Age. Chicago, IL: University of Chicago Press; 2009.
7. Wilson G. Researches on Colour-Blindness: With a Supplement on the Danger Attending the Present System of Railway and Marine Coloured Signals. Edinburgh, UK: Sutherland & Knox; 1855 (reprinted by TheClassics.us 2013).
8. Maxwell JC. Experiments on colour, as perceived by the eye, with remarks on colour blindness. Trans Roy Soc Edinburgh 1855; 21: 274–99.
9. Keeler R, Singh AD, Dua HS. “Wool over eyes”: Holmgren’s Skeins and Thomson’s Stick. Br J Ophthalmol 2009; 93: 32.
10. Edridge-Green FW. Two new tests for colour blindness. Brit Med J 1890; 1: 72.
11. Vingrys AJ, Cole BL. Origins of colour vision standards within the transport
industry. Ophthalmic Physiol Opt 1986; 6: 369–75.
12. International Maritime Organisation. Convention on the International Regulations for Preventing Collisions at Sea (COLREGs). London, UK: International Maritime Organisation; 1972.
13. International Civil Aviation Organisation. Annex 14 to the Convention on International Civil Aviation. Aerodrome Standards. Aerodrome Design and Operations. Montréal, Canada: International Civil Aviation Organisation; 2004.
14. Williams CH. Standards of form and color-vision required in railway service. Trans Am Ophthalmol Soc 1897; 8: 227–41.
15. Williams CH. Standards of form and color-vision in railway service. Boston Med Surg J 1897; 136: 561–6.
16. Williams CH. An improved lantern for testing color-perception. Trans Am Ophthalmol Soc 1900; 9: 192–6.
17. Williams CH. An improved lantern for testing color-perception. Trans Am Ophthalmol Soc 1903; 10: 186.2–9.
18. Cole BL, Vingrys AJ. A survey and evaluation of lantern tests
of color vision
. Am J Optom Physiol Opt 1982; 59: 346–74.
19. Taylor GH. Color testing and the psychology of color. Am J Psychol 1924; 35: 185–9.
20. Miner J. Colour blindness tests. Austral J Optom 1950; 33: 155–61.
21. Garner LF. Colour vision requirements in government departments. Austral J Optom 1976; 59: 369–75.
22. Taylor SH. The colour-blind face and voice. Br Med J 1910; 2608: 2005.
23. Edridge-Green FW. The expression of the colour blind. Br Med J 1910; 2609: 2051.
24. Taylor GH. Colour Blindness. Br Med J 1911; 2642: 410.
25. Mitchell LJC. The vision of railwaymen. Br J Ophthalmol 1922; 6: 319–21.
26. Martin LC. A standardized lantern for testing colour vision. Br J Ophthalmol 1939; 23: 1–20.
27. Martin LC. A standardized colour-vision testing lantern (II) Transport
type. Br J Ophthalmol 1943; 27: 255–9.
28. National Transport
Commission (Australia). National Standard for Health Assessment of Rail Safety Workers, vol 2. Assessment Procedures and Medical Criteria; 2004.
29. National Transport
Commission (Australia). National Standard for Health Assessment of Rail Safety Workers, vol 1. Management Systems. Incorporating the Guidelines for Health Risk Management; 2004.
30. Farnsworth D, Foreman P. Development and Trial of New London Navy Lantern as a Selection Test for Serviceable Color Vision
, Report No. 2. Groton, CT: Naval Submarine Medical Research Laboratory; 1946.
31. Paulson HM. The Performance of the Farnsworth Lantern at the Submarine Medical Research Laboratory and in the Field from 1955 to 1965. Groton, CT: Naval Submarine Medical Research Laboratory; 1966.
32. Paulson HM. Comparison of judgements by normals and color defectives to 408 colored stimuli. In: Verriest G, ed. Modern Problems in Ophthalmology. Colour Vision Deficiencies II. Basel, Switzerland: Karger; 1974: 221–4.
34. Cole BL, Lian KY, Lakkis C. Color vision
assessment: fail rates of two versions of the Farnsworth lantern test. Aviat Space Environ Med 2006; 77: 624–30.
35. Laxar KV, Wagner SL, Cotton TC. Evaluation of the Stereo Optical Co. Farnsworth Lantern (FALANT) color perception test: a specification and performance comparison with the original FALANT. Naval Submarine Medical Research Laboratory: Groton, CT; 1998.
36. Holmes JG, Wright WD. A new colour-perception lantern. Color Res Appl 1982; 7: 82–88.
37. Hovis JK, Oliphant D. Validity of the Holmes-Wright lantern as a color vision
test for the rail industry. Vision Res 1998; 38: 3487–91.
38. Hovis JK, Oliphant D. A lantern color vision
test for the rail industry. Am J Ind Med 2000; 38: 681–96.
39. Fletcher RJ. The Fletcher CAM lantern colour vision test: clinical results and calibration. Optom Today 2005; 45: 24–6.
40. National Transportation Safety Board (U.S.). Head-On Collision of Two Union Pacific Railroad Freight Trains Near Goodwell, Oklahoma, June 24, 2012. NTSB/RAR-13/02 PB2013-107679. Washington, DC: National Transportation Safety Board; 2013.
41. Rail Safety and Standards Board (U.K.). Annual Safety Performance Report 2012/13. London, UK: Rail Safety and Standards Board; 2013.
42. Independent Transport
Safety Regulator. Rail Industry Safety Report 2010-11. Sydney, Australia: Independent Transport
Safety Regulator; 2011.
43. Nikandros G, Tombs D. Measuring railway signals passed at danger. In: Cant T, ed. SCS ’07 Proceedings of the Twelfth Australian Workshop on Safety Critical Systems and Software and Safety-Related Programmable Systems—Vol. 86. Adelaide: Conferences in Research and Practice in Information Technology (CRPIT); 2007: 41–4.
44. Hall S. Hidden Dangers: Railway Safety in the Era of Privatisation. Shepperton, UK: Ian Allen Publishing; 1999.
45. Hall S. Beyond Hidden Dangers: Railway Safety into the 21st Century. Hersham, UK: Ian Allen Publishing; 2003.
46. Vingrys AJ, Cole BL. Are colour vision standards justified for the transport
industry? Ophthalmic Physiol Opt 1988; 8: 257–74.
47. Commission Internationale de l’Éclairage. Recommendations for Surface Colours for Visual Signalling. Vienna, Austria: Commission Internationale de l’Éclairage; 1983.
48. Casolin A, Katalinic PL, Yuen GS, Dain SJ. The RailCorp Lantern test. Occup Med (Lond) 2011; 61: 171–7.
49. Wood JM, Atchison DA, Chaparro A. When red lights look yellow. Invest Ophthalmol Vis Sci 2005; 46: 4348–52.
50. Gupta P, Guo H, Atchison DA, Zele AJ. Effect of optical aberrations on the color appearance of small defocused lights. J Opt Soc Am (A) 2010; 27: 960–7.
51. Hovis JK. When yellow lights look red: tinted sunglasses on the railroads. Optom Vis Sci 2011; 88: 327–33.
52. National Transport
Commission. National Standard for Health Assessment of Rail Safety Workers, 2nd ed. Melbourne, Australia: National Transport
53. Commission Internationale de l’Éclairage. Colours of signal lights. Vienna, Austria: CIE; 2001.
Keywords:© 2015 American Academy of Optometry
accident prevention; visual function; vision standard; medical standard; color vision; color-vision loss; color vision standards; railroad worker; transport; occupational risk; safety standard; lantern tests; practical test