In 1922 J. Owen Smith was a Radio Corporation of America engineer, and was also an official in the RCA-affiliated National Amateur Wireless Association. At this time Smith operated Special Amateur station 2ZL, located in Valley Stream, Long Island, New York, which he used to test and promote continuous-wave operations, on shortwave wavelengths below the traditional amateur wavelength of 200 meters. Smith's 2ZL was also one the stations heard in England during the historic December, 1921 trans-Atlantic tests.
 
Modern Radio Operation, J. O. Smith, 1922, pages 77-78:

CHAPTER  IX

Continuous  Wave  Transmission  by  Amateurs

    For a long time before the war the operators of amateur radio stations had discussed the possibility and practicability of the use of continuous wave transmitters for amateur work. Many theories, data and a few facts were submitted from time to time to prove that continuous wave transmission on short wave lengths was both possible and impossible.
    The advantages of continuous wave transmission, especially its economy and greater flexibility as compared to spark transmission, made its general use by amateurs highly desirable. The deterrent features were the impossibility of easily securing the means of generating undamped waves, and the important fact that continuous waves on short wave lengths were declared by many to be impractical, as it was believed that the slightest change in the characteristics of the transmitter, or the transmitting antenna system would cause audibility changes that would make successful reception impossible.
    After the ban on amateur transmitting had been lifted, a small supply of transmitting tubes became available and amateur experimenting with C.W. transmitting outfits started in earnest.
    Thousands of amateurs had used continuous wave transmitters on short wave lengths during the war in various branches of the service. Consequently had become more or less familiar with the general methods and results. As their actual experience with these sets had been more or less confined to attaching wires to binding posts on the outside of cabinets, they found considerable difficulty in securing and assembling the various parts and elements necessary. But it takes brains, energy and perseverance to be a successful amateur, and what might have proved a tough proposition to any other class or set of human beings didn't stop the determination of Young and Old America to possess a reliable, practical C.W. transmitting set. And so the work went on, causing more than one enthusiast to lose more hours of sleep than would possibly be sacrificed by the average person, that is, provided said average person desired to continue to live. A one amateur aptly expressed it, paraphrasing a well-known song--"The hours I spent with thee, dear set."
    But perseverance will generally win out sometime or other and the result of the untold hours of study and work on the part of the amateurs finally resulted in a finished product. A professional systematizer would undoubtedly suffer mental torture and anguish could he see the numberless types and specimens of C.W. outfits in use by amateurs today, for it would be practically impossible to find two that look as though they were even distantly related, but the important fact is, they work.
    Starting with the basic parts, a source of plate potential, a tube, or tubes, a home-made coil or two and what other miscellaneous junk could be borrowed or otherwise secured, these ambitious amateurs have always managed to get some amount of undamped energy into an antenna.
    Some of the circuits which have been tried and tried and tried were really wonderful creations. Some were plain and simple. In fact some were so simple as to be foolish. Others were complicated beyond description. The writer can readily recall his early attempts to secure reliable information on C.W. sets. To the best of his knowledge and belief, the crop totalled something like seventy-five different circuits, all of which were thrown into the waste basket with disgust after many weary hours of wasted trial and effort.
Pages 82-83:
THOUSAND-MILE  AMATEUR  RADIOPHONE

    The remarkable distance records made by amateur radio stations using continuous wave transmitters with several 5-watt tubes in multiple, are matters of common knowledge. The fact that a set, employing two or three of these small tubes, with approximately an ampere or more of current in the antenna, has been heard and worked by stations 1,000 or 2,000 miles away, causes no special interest at the present time. 2ZL photo
    Experiments at 2ZL station, Valley Stream, Long Island, covering several weeks, during 1920, with transmitting sets employing 5-watt tubes, clearly demonstrated that the signals from such sets are subject to practically the same conditions as damped transmitters, where the distance between the transmitting and the receiving station is more than the regular daylight range of the transmitting station, especially where the C.W. output is modulated in some manner. The foregoing, however, applies only to general conditions.
    On nights when the stations of the Eighth District, particularly those in Ohio, were inaudible on Long Island, 2ZL, using straight or modulated C.W., was also inaudible at several Eighth District stations, listening on a pre-arranged schedule. When the signals from Eighth District stations were audible on Long Island, the signals of 2ZL were copied in Ohio. During these tests one point of considerable importance developed. On nights when spark stations in the Eighth District "swung" so badly as to make consecutive reading of their signals impossible with two steps of audio-frequency amplification, the straight C.W. signals from 2ZL were reported as being good and steady, and consecutive reading was entirely possible. Summarized, this established the fact that, while the signals from 2ZL were subject to general conditions over long night distances when they were heard at all they were much more steady and reliable than spark signals. It should be remembered also, that the input of the tube transmitter at 2ZL was about 160 watts, plate and filament, as compared to 1,000 watts input--without counting the energy used to run the usual non-synchronous rotary spark gap--in the case of the Eighth District spark stations.
    After the experiments with the transmitter employing the 5-watt tubes had been carried on for several weeks at 2ZL, and the reports of the listening stations carefully studied, it became evident that the signals of such a small set were entirely satisfactory over distances up to 100 miles. Later on it was decided to increase the power of the set. It was not practical, of course, to use more than three or four 5-watt tubes, because the small added output of more tubes does not warrant the additional expense of the tubes or the extra filament and plate power.
    As the only obtainable tubes of increased size over 5 watts are of 50-watt output capacity, it was decided to install a transmitter employing tubes of that size, in accordance with figure 26. Right here is where the writer got into the same predicament as the man who caught a wildcat by the tail--he sure started a fine bunch of trouble for himself. It seemed logical to suppose that to install the 50-watt tubes it would only be necessary to insert the new tubes and sockets, supply proper filament and plate voltage and shoot the moon. But it was somewhat different before a smooth working 100-watt set was finally developed.
    The characteristics of the larger tubes were such that their output was 50 watts on 1,000 volts plate potential. Of course, it was decided, amateur fashion, to get every single possible watt out of the tubes.
Pages 87-89:

    The signals of 2ZL were reported from stations at the following points while using the set just described and 2ZL worked with a number of them:
Miles--
Air Line
from 2ZL		     Miles--
Air Line
from 2ZL
Palm Beach, Fla.900 New Orleans, La.1,079
Stanbridge East, Quebec      350 St. Paul, Minn.1,025
Little Rock, Ark.1,025 Kansas City, Mo.1,080
Chicago, Ill.725 Louisburg, Nova Scotia      550
Detroit, Mich.500 Montreal, Quebec350
St. Louis, Mo.850 Marion, Mass.220
Houston, Tex.1,286 Memphis, Tenn.925
Port Arthur, Tex.1,265 Grand Forks, N. D.1,300
Ellendale, N. D.1,330 Minneapolis, Minn.1,025

    Using one tube as an oscillator and one as modulator the voice has been reported from Marion, Mass.; Boston, Mass.; Memphis, Tenn.; Anderson, Ind.; Niles, O.; Washington, D. C.; Rochester, N. Y.; Pittsburgh, Pa.; Cambridge Springs, Pa.; and Montreal, Quebec, and many other points, and two-way communication has repeatedly been carried through successfully; 2ZL by voice and the other stations replying by spark.
    The work done when using straight or unmodulated C.W., established new records and distances for amateur C.W. transmission, at that time, in that daily schedules were maintained with the following points: Savannah, Ga. (4XB); Salem, Ohio (8ZG); Canton, Ohio (8ZV); Langley Field, Va. (XF-1). Occasional communication was established with Madison, Wis. (9XM), and Minneapolis, Minn. (9XI).
    On one or two occasions when communication had been established with Western stations, it developed that no Eastern amateur spark stations had been heard at Western stations, and no Western ones at 2ZL. The straight C.W., however, was going through without any trouble. Slight fading was noticeable, but of a longer period and more gradual character than in the case of spark signals.
    Insofar as the dependable daylight range of the set was concerned, distances varied with the method of transmission or modulation. Using straight C.W., 2ZL was copied repeatedly during daylight over varying distances up to 200 miles, the greatest distance having been Boston, Mass. Conclusive tests to determine the maximum C.W. daylight range were not made, but it is believed that with three amperes of straight C.W. in the antenna it should be possible to communicate with stations 300 miles distant during fairly favorable daylight conditions.
    With the output of the set at 2ZL modulated by buzzer, daylight communication has been successfully carried on with stations 150 miles distant, although the reception, of course, called for careful tuning and generally much preliminary transmitting for adjustment of the receiving sets. With voice modulation, during daylight, conversation with stations seventy-five miles away has also been successfully carried on. When the two tubes were used in multiple, with the tone wheel chopping the grid-leak current, the received signals were several times the audibility of buzzer or voice modulation.
    As a summary, the dependable daylight ranges of a set with approximately three amperes in the antenna, can be reasonably assumed to be as follows:
Straight C.W.200 miles
Buzzer modulated75 miles
Voice75 miles
Tone wheel chopper   100 miles

    When these ranges and the flexibility of the set are considered, and also the fact that the total input of the set, plate and filament, under all conditions of transmission never exceeded 350 watts, the very great advantage of C.W. transmission over the usual spark method is readily apparent.
    As a side light on the voice transmission tests made at 2ZL, at that time a letter was received from an amateur in a little town with a population of 160, located thirty miles South of Memphis, Tenn., saying that he frequently heard the phone at 2ZL station. He stated that he used a small aerial and only a single tube as a detector. He said that he and his people liked to listen to the phone and requested that the voice be used often, that music be played occasionally as he and his people enjoyed listening to it.
    Wonderful and mysterious are the ways of C.W. transmission when an amateur in Tennessee, 925 miles distant, regards a radiophone concert for his benefit by another amateur station in the vicinity of New York as an ordinary matter.
    With the set previously described all then known amateur C.W. transmission records were broken by 2ZL and 5XB stations on February 11th and 12th, 1921. The latter is the station of the Agricultural and Mechanical College of Texas, located at College Station, Texas. The transmitter used there consists of three 5-watt tubes, the total input plate and filament being approximately 175 watts. The overland air line distance between the two points is 1,500 miles. The two stations were in communication occasionally during the winter of 1920-21 and messages were exchanged successfully in both directions.
Pages 91-94:

    Some experiments were made at 2ZL to determine the practicability of employing wave lengths below 200 meters in connection with tube work. A separate antenna, considerably smaller than the main antenna regularly used, was used for this short wave work. This smaller antenna was about 60 feet long overall and consisted of four wires. Considerable work was done on 175 meters. The antenna current on this wave length being two amperes, it was found entirely possible to work 100 miles in daylight on this wave length without trouble with that amount of current in the antenna. The antenna current on 150 meters was in the neighborhood of 1½ amperes. As practically no amateur stations were equipped with receiving apparatus which would accommodate a wave length of 150 meters, however, it was found impossible to make any experiments on this wave length to determine the daylight range of the set. When the transmitter was adjusted to a wave length of 175 meters it was found, in at least three instances, that the receiving operators had to adjust their secondary circuit variometers at zero in order to hear the signals. When the wave was further reduced it was found impossible to "raise" any of the listening stations. After communication had been carried on for some time on 175 meters considerable comment was made by other amateur stations on the desirability of working on that wave length because of the absence of interference and very little trouble was experienced from atmospheric disturbances on nights when static was giving considerable trouble on wave lengths above 200 meters. A great deal has been heard from various points to the effect that it is difficult or impossible to secure sufficient antenna current on a wave length of 200 meters to enable the transmitting station to work any respectable distances. It would seem that this condition is due entirely to the fact that many amateurs attempt to adjust C.W. outfits to a 200-meter wave on an antenna with a fundamental wave length of 200 meters, which arrangement, of course, precludes sufficient coupling in the tuning arrangement to allow free oscillation of the set. It is, however, entirely possible to work on 175 meters with tube transmitters, without trouble, providing the antenna system is of the proper size for that wave length. The idea that tubes will not operate and generate power on 200 meters is an absolute fallacy which has evidently arisen through lack of knowledge, or because of misinformation. Tubes will oscillate on short wave lengths just as well as on long wave lengths, providing an antenna of the proper characteristics is used.

THE  AMATEUR  TRANS-ATLANTIC  TRANSMISSIONS  OF  DECEMBER,  1921

    The story of the amateur trans-Atlantic tests of December, 1921, is so well known, that it seems unnecessary to tell more than a condensed story of the unprecedented accomplishment in these pages. At the first national convention of the American Radio Relay League, held at Chicago, during the summer of 1921, it was decided to send Paul F. Godley to Europe, with proper receiving equipment, to determine if the signals of American amateur stations could be heard across the ocean, an approximate distance of 3,000 miles. Preliminary trials were held during November and all amateurs were invited to participate, with the understanding that 1,000 miles must be covered to make any station eligible to enter the trans-Atlantic trials on an individual basis. Twenty-seven stations qualified and to each one a group of code letters was assigned, as well as a definite period of transmission. Free-for-all periods for each district were also set apart each night and all stations were invited to participate.
    The result was not at all what had been expected. Many of the stations which had qualified in the 1,000-mile preliminaries were not heard on the other side at all, while several stations which failed to qualify or which had not participated in the preliminaries were heard, either by Mr. Godley or by English and Dutch amateurs. One station, 2AJW, heard by Godley, employed 5-watt tubes, the input being approximately thirty watts. Several other stations which were not heard at all by Mr. Godley were copied by English amateurs and the code-words verified, precluding any error in reception.
    A complete list of stations heard by Mr. Godley at Adrossan, Scotland, is as follows:

Spark:
1AAW, not yet located;         3FB, Atlantic City, N. J.
1ARY, Burlington, Vt. 8BU, Cleveland, Ohio;
1BDT, Atlantic, Mass. 9ZJ, Indianapolis, Ind.
2BK and 2DN, Yonkers, N. Y. 3BP, Newmarket, Ontario.
2EL, Freeport, L. I.
Continuous Wave:
1ARY, Burlington, Vt. 2ARY, Brooklyn, N. Y.
1BCG, Greenwich, Conn. 2AJW, Babylon, L. I.
1BDT, Atlantic, Mass. 2BML, Riverhead, L. I.
1BGF, Hartford, Conn. 2EH, Riverhead, L. I.
1BKA, Glenbrook, Conn. 2FD, New York City
1RU, Hartford, Conn. 2FP, Brooklyn, N. Y.
1RZ, Ridgefield, Conn. 3DH, Princeton, N. J.
1XM, Cambridge, Mass. 2ACF, Washington, Pa.
1YK, Worcester, Mass. 8XV, Pittsburgh, Pa.

with the probability that 4GL, Savannah, Ga., was also heard.
    During the tests the following American stations were heard by English amateurs:
    1BCG, Greenwich, Conn.
    1AFV, Salem, Mass.
    1UN, Manchester, Mass.
    1XM, Cambridge, Mass.
    1ZE, Marion, Mass.
    2FP, Brooklyn, N. Y.
    2BML, Riverhead, Long Island.
    2ZL, Valley Stream, Long Island.

    All the stations heard by the English amateurs used C.W. transmitters during the tests. Practically every type of circuit was employed, as 1BCG used D.C. on the plates, 1ZE Kenotron-rectified, 60-cycle A.C., 2FP 500-cycle A.C., 2BML half-wave rectification of 60 cycle A.C., and 2ZL full-wave rectification, 60-cycle A.C.
    In commenting on the result of the tests. Mr. Godley made the following statement:
    "In glancing over the above lists one is struck by the preponderance of the C.W. stations, and by the fact that the British heard C.W. stations only. That can mean only one thing, that C.W. is far superior, and I should like nothing better than to see all amateurs change over to continuous wave at once. Spark methods are horribly out of date and are so inefficient, comparatively, as to be ridiculous, were it not that many have invested good money in spark equipment. Station 1AFV, since the tests, has gotten three messages across to England (London) on 200 watts of C.W. Many stations of the Atlantic seaboard are reaching to the California coast with similar powers, while the west coast stations have been shoving signals into the Hawaiian Islands. The day is not far distant when amateurs the world over will be exchanging greetings in many languages, and by the same token, the day is almost here when spark stations will be of interest as having to do with history only."
    The set used at 2ZL station at Valley Stream, L. I. when signals from the station were heard in England, employed two 250-watt Radiotrons, UV-204, in a full-wave rectification circuit.