In addition, Mercury elongations vary in favorability. Indeed, for mid-northern latitudes, spring evening (eastern) elongations are the most favorable, with respect to the maximum angular distance from the sun, length of visibility (usually less than about 20 days) and height above the horizon. The same is true for the autumn morning (western) elongations; the entire situation is reversed for the southern hemisphere. The primary factor controlling favorability is the inclination of the ecliptic to the horizon.
The earliest observations of Mercurian detail were those of cusp truncation by Schroter, in 1800. Occasional truncations were observed by others but little surface detail was recorded until the 1880's, by Denning in England and Schiaparelli at Milan. Schiaparelli published no individual drawings to this writer's knowledge but did publish the first planisphere (single hemisphere) of Mercury's surface (1889). He believed the planet to rotate synchronously with a period equal to its orbital period, namely 87.969 days. This opinion was not at first universally accepted. Schroter's assistant, Harding, and the astronomer Prince concluded that the Mercurian day was approximately that of the Earth's (Sandner, 1963). Leo Brenner (1896) concluded that it was actually 33 to 35 hours. These opinions lapsed after various observers would observe a feature as motionless over a period of several hours (of date-time observing)(see, for example, McEwen, 1909).
The first third of this century found the observations of Mercury confined principally to three groups:
The most important work of the period is that of Antoniadi (1934).
Working at Meudon, Antoniadi observed Mercury from 1924 to 1929, firmly
established (as he thought) an 88--day rotation period, and gave
nomenclature to the various bright and dark areas. Although not the
first to do so, he had truly mastered the art of daylight observations.
The drawings, being so excellent, if small in number, firmly entrenched
various markings into Mercury observation lore, such as his famous
"figure 5" (See Sandner, 1963, page 36), so much so that observational
bias imitative of Antoniadi is quite strong in many later drawings.
Mercury observation became sporadic after Antoniadi's work. Only a rather well-observed elongation in 1936 (McEwen, 1936) was reported in the 1930's plus a few French contributions (Danjon 1937, Quenisset 1931). In 1942 and 1944 Lyot and Camichel at Pic du Midi Observatory observed Mercury and made the first good photographic series of observations (see Dollfus, 1953). In 1950, Dollfus also obtained a very fine series of drawings (1953) using filters for increasing the visibility of features. The French efforts effectively ended, in terms of published material, at this time and the JBAA has published a scant nine observations for the 1950's and 1960's.
Most of the post-World War II drawings are to be found in the American journal, The Strolling Astronomer, of the Association of the Lunar and Planetary Observers. Created by Walter Haas, who originally published in the JBAA, this journal also has a Mercury section, which publishes selected drawings for each year. Unfortunately, many, if not most, of the drawings are inaccurate and unusable.
A short discussion of the Mercury observations is appropriate here. As noted previously, the markings are at the limit of visual perception, due somewhat to their own soft nature but more to the observing constraints. The personal equations (personal styles and tendancies) of the observers are quite significant. Often, general features agree between observers observing the same apparition but the detail and placement of the markings often show little, if any, agreement (see, for example, the drawings of the Fournier brothers (Jarry-Desloges, 1910)). Once the 88-day rotation period was "confirmed" by Antoniadi, it became rather easy for an observer to "see" Antoniadi's markings in the fleeting images of steady seeing.
Still, discrepancies were reported from time to time. Major dark areas
disappeared, bright areas of the disk and limb would appear, even polar caps
were recorded on both poles. Some Mercury observers, especially Haas (1947),
ascribed these occurences to clouds and haze in a Mercurian atmosphere. Most of
such phenomena may be ascribed to the personal equation of the observer (where
or not he was Antoniadi-biased), others imply poor observing and/or poor
observer. Some short-lived phenomena have been ascribed possibly to luminescence
(Cruikshank 1966), but the aforementioned factors may still play a role. Any
apparent motion of the features was considered as librations, a rocking of about
23 degrees amplitude in longitude and about seven degrees in latitude. But the
major cause of discrepancies is the true nonsynchronous rotation of Mercury, so
that observers were really viewing more than just the one hemisphere they
thought they were seeing!
The 88-day rotation period seemed quite secure, both observationally and theoretically. Tidal friction from the sun's pull caused Mercury to have one face eternally sunward, a situation similar to the Earth-Moon system (so it was thought). It, therefore, was a considerable jolt when Pettengill and Dyce (1965), using radar doppler techniques, found that they couuld fit their measures only to a 59 +/- 5 day Mercurian "day." Peale and Gold (1965) proposed immediately that Mercury was locked into a exact 2/3 resonance with the orbital period. Other astronomers quickly proved the strong resonance theoretically (Columbo 1965, Columbo and Shapiro 1966), which immediately created new consequences. The motion of the Sun in the Mercurian sky caused the formation of "hot poles" 180 degrees apart (Soter and Ulrichs (1967)) and reinterpretation of the infrared data was necessary to evaluate light and dark side temperatures (Morrison 1970, Section 7.1). *Our* most important concern would be how much of the old visual observations are still accurate and useful. Indeed the question "Why was the 59-day period not discovered visually?" immediately begs an answer.
The answer lies in several interesting coincidences. A point on Mercury at a noon position will be at the noon position at the same orbital position exactly 3 Mercurian days later, which is also exactly 2 Mercurian years. Now, the synodic (phases) period is 116 days, which is 4/3 the orbital period, twice the Mercurian day, and about one-third of an Earth year. The synodic period means that there are 116 days between similar apparitions but not all of these are favorable in mid-northern latitudes, where the observers were. Favorable apparitions tend to occur at every third apparition.
This situation, for a particular face of Mercury, will exist for six years after which the favorable apparitions will shift by 1 synodic period, for the synodic period is slightly out of phase with the calendar year.
Thus, if an observer examined Mercury for just the favorable apparitions for a few years, he would always see the same features which leads to asynchronous solution!
There is one more coincidence to the puzzle. Maps made by the best observers, computer digitized and placed upon a map based on a 59 day rotation, still show the famous "Figure 5", a set of dark markings that resemble the numeral. The problem is....the "5" is replicated in three places! Though one is by far the most prominent, the three together, each 120 degrees in hermographic longitude apart mean that once more, an observer could be fooled into believing he has seen the same feature!