Almost 30 years ago, in 1978, the astrological world was changing fast. A clear sign of this shift in momentum: an astrologer named Michael Erlewine had just started publishing Matrix Magazine. This skinny technical journal was a treasure trove of the magic program code that made astrological computing possible on small computers and programmable calculators. A lot of this early code was written under the pen name James Neely, by a man who had both the mathematical and astronomical expertise to understand the formulas and equations behind the calculations.

This new interview with James Neely is an opportunity to recall a time that was seminal for our generation of astrologers, a whole Saturn return ago. It also offers quite a novel glimpse of the man himself and of his more recent pursuits. Neely was a pioneer in the development of personal astrological computing, and for his unique contributions he received the Johndro Award that was presented in 1978 by several participating journals: The Journal of Geocosmic Research, Matrix Magazine, Cosmecology Bulletin, Phenomena, Sirius, and the CCRS Newsletter.

Since that time, Neely has been instrumental in other significant advances in the field of astrological computing, including his research on ephemeris calculations for Chiron and for the Uranian planets. His papers on accuracy of planetary calculations, determination of the times of planet stations, and the mathematics of house computations in extreme latitudes are all classics in this field.

An Interview With James Neely

By Kyle Pierce

When did you become interested in astrology, and what did you study?

For the record, I am not an astrologer and make no claims of being one. However, I do know a fair amount about the mathematics of astrology. Exactly how this all came about, I'm not quite sure. During the late ‘60s and early ‘70s my wife and I became quite interested in things metaphysical. For a number of years we sponsored a weekly "spook group" in our home. This included a well-known deep trance medium, who was an occasional house guest, and a number of local people whose interests ran to table tipping, psychometry, a one-person Ouija board operator, handwriting analysis, and, of course, astrology.

Some kept telling me I should get into astrology because of my interest in mathematics. I did attempt to learn astrology and studied with the Rosicrucian Fellowship for several years, but this never did really take, as the rote memory drove me up the wall. The thing that really fascinated me was the construction of a chart. I think this was because I always have had an interest in astronomy and being able to calculate the position of the planets was intriguing.

My interest really picked up when the Hewlett Packard HP-35 handheld calculator came on the market in 1972. I recall that previously I had spent some 2 1/2 to 3 hours solving Kepler's equation using tables of logarithms and trig functions. When I got my HP-35 I went back to the same problem and, with just the keyboard and one memory location, solved it in 15 minutes. Actually, I didn't have to wait for the handhelds — as it turned out my computer interests started when the first version of Fortran came out in 1956, but I couldn't justify using the company mainframe for personal interest calculations, so the planetary stuff sat on the back burner until the advent of the programmable handhelds.

When did you first get a programmable calculator up and running with a planetary routine? This would be a landmark event, I think. I got a TI-59
and entered all your code from Matrix Magazine… maybe in ’78, but it
might have been later.

Looking through my archives I see that my interest in astrological calculations started in 1973 when I began looking at the various house systems and redoing the mathematics in a way that was understandable to me.

The next year the first programmable handheld calculator, the HP-65, priced
just under $800, was a marvel for its time. It had 100 program steps and 9 storage registers and was ideally suited for programming the house cusp mathematics.

About the same time, HP introduced its Users' Library of submitted algorithms. I called Hewlett-Packard and asked if they would accept astrology programs for their library. The answer was "yes." That was when James Neely was born. Under that name I submitted 44 HP-65 astrological programs to the Users' Library. The early submissions were, for the most part, programs for various house systems. These were easy because they could be put on one magnetic card of 100 or less program steps. Then I discovered that by overlaying a program I could compute the rectangular coordinates of the Sun with three overlays, then three more overlays to get the heliocentric rectangular positions of a planet. From there the ecliptic and equatorial position could be computed. Later the four major asteroids were added to the collection along with aspects, midpoints, and distance values.

Any memories or observations you wish to share about that early pioneering stage of astrological computing? Your contributions, Michael Erlewine, Matrix Magazine, other contributors?

This was an interesting time because suddenly I had all sorts of pen-pals, including Buz Overbeck, who introduced me to sidereal astrology and a variety of other things which could be calculated. As a result, the planetary programs offered both tropical and sidereal calculations, and house programs were extended to include mundoscopes, octoscopes, and the like.

Then in July of 1976, Hewlett-Packard introduced the HP-67 and its printing cousin, the HP-97. This opened up a whole new world because now we had 224 program steps available as well as 26 storage registers, so re-programming began. HP expanded its Users' Library to include the new machines. Then they began printing collections of users' contributions on similar subjects, such as Calendars, Space Science, Navigation, Surveying, Physics, Mathematics, etc., and Astrology. Anxiously, I opened the Astrology collection to see what it contained: there were 11 programs in all. Ten of the programs (handwritten) were by James Neely and one program (neatly typed) was by some interloper named Michael Erlewine. I looked over his program to see what it contained. To my mind it was all well done except that he included a routine to sort computed angles for the proper quadrant. This gave me an excuse to write to him, find out who he was, and tell him there was a direct way to put the angle in the proper quadrant if he was interested.

Thus a long series of correspondence and interactions started. Michael sent
me an early Commodore PET 2001, which was introduced in 1977. I translated most of the HP programs into Basic on the PET and Michael and Stephen Erlewine incorporated many of them into the early Matrix software astrological program. The planetary algorithms underwent a number of improvements in accuracy, from about a degree of arc for the very early programs to within a few seconds of arc as more accurate algorithms became available. Today, thanks to the Jet Propulsion Laboratory (JPL), the accuracy is mind boggling. JPL states that their lower accuracy ephemeris, covering the years -3000 to +3000, is no worse than 25 meters for any planet and no worse than 1 meter for the Moon! They use their high accuracy ephemeris for planetary mission planning.

As Matrix was building and improving their software line, they also published Matrix Magazine, the first appearing as the Winter 1978 issue. A good portion of my work was presented in this series, including the algorithms for the handheld calculators, both Hewlett-Packard and Texas Instruments.

Occasionally all of this had a light side. One incident in particular involved an article that was published in the AFA Bulletin. A well-known astrologer had written an article about someone she knew. I don't remember any of the details of the article except that it struck me as very cutesy the way she was trying to hide his identity and everything about him except what she was expounding upon. And, of course, she had published his chart from which she could make her various points. I thought, "that's ridiculous, lady, you're trying to hide everything about him but yet you've published his chart!" To me, this meant she had told us everything except his name. Working from the positions of some of the slower moving planets inwards, it was no problem to determine the year of birth. The solar position revealed the date of birth and the lunar position determined the time of birth. The position of the MC gave up the longitude and the ASC gave the latitude. With this, the atlas gave up the particular city. I wrote all of this to the editor of the Bulletin and he liked it. He was going to publish it, but he sent a copy on to the astrologer. Talk about angry! She accused me of malpractice and all sorts of things. All I could do was tell her that she was the one who let the cat out of the bag by publishing his chart. I also told her I could have been a bit more accurate if I knew whether she used a noon ephemeris or a midnight ephemeris.

One other little tidbit I had hoped to publish in detail one day, but never got around to it, is a relationship between the Placidus and Topocentric house systems. Actually I put the method into Program 01646A of the HP-65 Users' Library which gives three options for approximations to the Placidus house system. The gist of the matter is this. In a 1968 publication ["Mechanics of Tables of Houses," Golden Seal Research Headquarters, Hollywood, CA], J. Allen Jones outlines approximation methods for computing Placidus cusps. At the time, approximation methods were all we had, for the exact solution to Placidus cusps requires an iterative process since there is no closed form solution to the equations. All Placidian approximation methods are based upon a spatial semi-arc trisection which approximates the temporal semi-arc trisection used in the exact method. The Raphael tables are computed from a trisection of the semi-arc whose declination is equal to the obliquity of the ecliptic (about 26°26'), which is the declination of 0° Cancer. The Dalton tables are computed from a trisection of the semi-arc whose declination is 18°16', which is the declination of 22° Taurus. Other approximations could be made by using any other reasonable declination. But if you take the limit as the declination approaches 0°, which is the declination of 0° Aries, the results are identical to the Topocentric system of Polich and Page. I wrote to Polich about this and he was delighted to learn of the correspondence between his system and the Placidus system.

Your work on orbital elements for TNPs has long been the standard source, I believe. For instance, in the Swiss Ephemeris source code there is a comment about “James Neely's revised elements of Uranian planets” along with the code that makes use of your formulas for the Uranians. How did you get involved in the problem of computing the Uranians/TNPs?

In 1977 I wrote five short articles relating to some planetary positions and earthquakes, these appeared in Phenomena, published by Malcolm Dean in Toronto.

One major piece of work was to examine the ephemerides of the transneptunian planets of Witte and Sieggrün to determine whether they obey the Newtonian laws of physical bodies or if they are combinations of cycles of known bodies. Although at the time there were several published ephemerides of the TNPs, they had to be considered as secondary or derived sources and so their accuracy was suspect. Copies of the original ephemerides
of Witte and Sieggrün were obtained and were the sole sources examined. With the aid of the Commodore PET, valid orbital elements for the TNPs were obtained. From the tables of Witte it was also possible to derive latitudes for two of his four TNPs. The tables of Sieggrün's four TNPs yielded valid orbital elements but are suspect as real bodies for there was no available latitude data and the orbital eccentricities are all zero (circular orbits). Nevertheless, these analyses of the primary sources did yield valid orbital elements which could be used for the computation of improved ephemerides. The analyses of the Witte and Sieggrün ephemerides were published in the Journal of Geocosmic Research, Vol. 2, No. 2, 1978, and the orbital elements of the TNPs appeared in Volume VII of Matrix Magazine.

Another project was the computation of an ephemeris of Chiron from the orbital elements published by Brian Marsden of the Minor Planet Center at the Smithsonian Astrophysical Observatory. This work was done in conjunction with Malcolm Dean in 1978.

Additional articles appeared from time to time in Matrix Magazine and later in Matrix Journal, ending with a highly mathematical article in 1992 involving great circle intersections and their relationship to sensitive points.

At least two other major pieces of programming were completed. One was to implement the “Low Precision Formulas for Planetary Positions,” by Van Flandern and Pulkkinen of the U.S. Naval Observatory. By low precision, the authors claim the formulas are good to one minute of arc for any epoch within 300 years of the present. Actually it was found that, when corrected for nutation and aberration, the formulas were accurate to at least 0.2 minutes of arc.