Voynich star names: an analysis (1)



The Voynich manuscript (VM) contains dozens of pictures of stars alongside what appear to be labels and names.

In our attempt to decipher the manuscript’s content it would therefore seem an obvious strategy to look systematically at these labels for clues to help us in decipherment. We could draw on the extensive research and scholarship into the star names we use today, and the origins, transmission and transmutation of those names over centuries.

This would help in decipherment because if we could find common star names in the VM and in systems which we know from elsewhere, this could give us important clues as to the sound-sign correspondences in the VM which could help us to crack the code This is in analogy with what I have tried to do in my earlier paper with (mainly) plant names.

In this light I find it remarkable that in the 100 years or so since the VM’s rediscovery there has been almost no serious systematic study of these star names. By ‘serious and systematic’ I do not include those attempts which choose possible names to support a pre-formed argument or theory (such as the approach used by Han here). I am thinking instead of more academic studies which would start with an open mind and seek to look at the evidence of the star names in the manuscript as it stands, and then compare the evidence with what we know about other star naming systems, without any grand a priori assumptions, and not seeking to support a pre-formed conclusion about the manuscript.

You can see for yourself how rare (i.e. non-existent) such studies are by means of a search through Google Scholar, as I have done here. Apart from many links which are irrelevant because they don’t consider the VM at all, such a search brings up remarkably few relevant discussions. The few which do emerge are d’Imperio’s study (1978), to be discussed below, a few papers by Brumbaugh (e.g. 1976) also to be considered below, a link to the Montemurro et al.  paper which hardly mentions stars at all, and nothing else of substance at all. That’s surprising, isn’t it?

In this and some later blog postings I hope to make a start to remedy that deficiency. It is obviously not possible in this kind of forum to produce the full, professional analysis of VM star names which I think is both necessary and possible, but I intend at least to make a start .

One of my aims, as in this website as a whole, is to open up discussion to other people – so feel free to add your views as we go.

Before we start on the discussion itself however, let us look at those few studies which have been written, i.e.those mentioned above.

D’Imperio’s study

D’Imperio’s 1978 work, which is still required reading by anyone interested in serious study of the manuscript (and which can be found in full here) offers no real analysis of the star names other than to say:

“Attempts to cross-match the rings of text around sun and moon, or the labels of individual stars on the two folios have so far been fruitless.” (D’Imperio p17)

However,  she clearly feels it is worth studying, since she says of the pages with star and other cosmological information:

“[s]urely a careful and determined analysis of this wealth of structured content in conjunction with a study of medieval doctrines should turn up something of use to us in interpreting the meaning of the diagrams.” (p17)

I agree with her completely on that. She feels that such a study had not been done up to 1978; also interesting is her view that such a study should be linked with what we already know about mediaeval ‘doctrines’. What is surprising to me is that no-one since 1978 has tried to do it either, at least in terms of the star names.


In the 1970’s (in fact before D’Imperio’s monograph) Brumbaugh produced a series of claims that he had ‘cracked the code’, among them an article on the VM star names  which he claimed he had deciphered (see the article here). In that article he boldly states: “In 1972 I finally solved the Voynich cipher” (Brumbaugh 1976:140).

Before addressing his decipherments themselves, it is interesting to reproduce a few of Brumbaugh’s observations concerning what he terms the ‘star names’, to give a flavour of his work:

“The star-map section raises problems. The star maps are a set of twelve, representing ten signs of the zodiac; originally there must have been at least one for each sign of the zodiac, but if so, two have been lost. Each ‘map’ has a central medallion with a picture of the zodiacal constellation, with the month name written beneath in plain text. Around this, in an inner circle, the maps have a ring of female figures, each attached by a line to a star. An outer circle has more such figures. A group of numerals is written next to each star.” (Brumbaugh 1976:141)

On the next page he continues as follows:

“A group of cipher numerals is written next to each star. A first try at decipherment gave the name of Alfred for the star in the Pisces medallion, with Urifydes just above and Alfansus Purus on to the right. The names are therefore those of the great men whose souls are attached to stars, rather than anything else they might be—Arabic star names, or a numerical Messier listing, for example.” (Brumbaugh 1976:142)

Brumbaugh does not give details about his decipherment, and it is even not clear from this which precise words he was deciphering (e.g. in the Pisces ‘medallion’). Unfortunately, partly because of this sort of vagueness, his resulting decipherment did not find favour with other scholars. D’Imperio was carefully polite about it, but concluded that:

“There is just enough plausibility in the process to lead one on. but not enough to leave one satisfied.” (D’Imperio 1978 p38)

and his scheme of decipherment is now largely forgotten.

In terms of his analysis of the star names, I would further argue that he made a number of fundamental assumptions which are questionable, even untenable. In the quotations  above, for example, he assumes that what we now call the Zodiac pages are in fact ‘star maps’. This involves the assumption that the drawings of star-shaped objects attached to lines or strings are necessarily meant to represent stars, as he says on pages 140 and 141, although by page 142 he has revised this to suggest they represent not stars themselves but ‘great men whose souls are attached to stars‘.

This raises a crucial question which to my mind (and again to my surprise) has not been carefully addressed in Voynich scholarship so far, neither by D’Imperio, nor by Brumbaugh, nor apparently by anyone else, namely which star-like drawings with labels in the manuscript should we take with confidence to represent stars, and which should we exclude from our analysis because they might not represent stars? According to page 18 of Tiltmans’ 1968 article, there are about 550 “star shapes” in the Voynich manuscript, but these cannot all represent actual stars. We need a principled discussion to determine which to include in our analysis and which to exclude.

To put this another way, if we are to study the star names, we need to establish from the outset a principled ‘corpus’ of items of those within the manuscript which can then form a solid basis for our subsequent analysis. This corpus must exclude items which might not be star names at all, or which are in any way doubtful. Without that clear and confident corpus, selected, argued and justified, we cannot be sure we are talking about star names at all.

So this first post will try to bring some clarity and rigour to this important question. My aim will be to determine a clear and defensible list of those ‘labels’ or ‘names’ in the Voynich manuscript which are most likely to represent actual names of stars, so than we then have a solid ‘corpus’ on which to work in future.

Star-shapes with lines or strings

f104r_bottomLet us start by considering star-shapes with strings, such as these (illustrated) from the bottom of f104r, and also those in the ‘Zodiac pages’ such as these ones . Some writers are rightly cautious about these, calling them “flower-like star shapes” (Montemurro et al.). Others are less cautious; René Zandbergen on his pages speaks of  some of them as “the drawing of a star in the margin” and also “30 nymphs holding stars”.

In my view we need to very cautious about these, since it is by no means certain that in fact they represent stars at all. They could be meant to be flowers, or suns, or something else. In their context on the Zodiac pages they could well represent ‘days’, since there is one for each of the (approximately) 28 figures in most of the months. In the last pages of the manuscript, although some of them do not have strings, such as those on page f103r here, they again seem to number around 360 or so, which again suggests that they might be intended to indicate days. In short, whatever they do represent, it is by no means certain that they are intended to represent stars, so we should exclude them.

This means that it seems more rigorous, and safest, to exclude all of these star-shapes on those pages, with or without strings, from consideration as we attempt to set up a list of star names in the Voynich manuscript, so I will not include them here.

Star-shapes in the ‘Cosmology pages’

In the co-called ‘Cosmology pages’ there are also many star-shapes, some with interesting labels beside or near them, for example on folio 67r1 here  or 68v3 here. However, in these cases it is again doubtful whether the labels are definitely meant to represent the names of the stars. Owing to the uncertainty over the meaning of the diagrams as a whole, and the positioning of the labels, they could well be intended for something else, so again it is wise to exclude these from consideration as we try to build a corpus of potential star names in the VM.

Folios 68r 1-3

However, on three of the folios of the manuscript we do seem to have a set of star shapes with associated labels which seem most probably to represent the names of the stars they are associated with. The reasons are as follows:

  1. on these pages the stars are clearly separate from each other, with only one label per star
  2. since they have a moon and a sun on two of the pages, and a moon on the third, they must be stars and not (say) flowers or days
  3. in the first two folios (f681-2) the star illustrations seem clearly to be positioned in the sky (between the moon and sun), and in the third they surround the moon, so again they must be meant for stars
  4. the labels beside each star are all different from each other, again making it likely that they are names

These elements of the context make it highly likely that the labels besides the stars in these folios are intended as names for the stars themselves. Indeed there does not seem to be any other plausible explanation of these labels in these positions.


f68r3 – the seven stars of Taurus

Even so, we still need to be cautious; in at least one case, namely the set of seven stars on 68r3 (illustrated), it has already been argued (e.g. in my Feb 2014 paper here) that this label probably represents not the name of a star or stars, but the constellation Taurus. We also know that many star lists in mediaeval times included the planets alongside stars (see discussion here) , so among star names in the Voynich we might well expect to find names for Mercury, Venus, Mars, Jupiter and Saturn. So when we say ‘star names’ in this discussion we need to remember than we might in fact also be dealing with constellation names and planet names too.

Notwithstanding, it seems incontrovertible that the labels on these three folios must be intended as names for the objects they stand beside, be they stars, asterisms or planets. We will therefore use the labels on these three folios pages as the basis of the corpus or list of Voynich names which we can then use to compare with other star name lists in other cultures.


Numbering the stars 

In order to discuss and attempt to analyse the stars and their possible names, it seems important to number them so that we can discuss them more easily and specifically. To this end I propose to use the following three illustrations of Folios 681-3, with a number beside each named star. These numbers will then allow us to list and analyse the star names more systematically – there are 64 in total.




Here I offer a download link so you can get a Word file with a list of these numbers and names.

Download (DOCX, 6.17MB)

I will also put up lists of the star names on separate pages for each folio:

Folio 68r1 Star names

Folio 68r2 Star names

Folio 68r3 Pleiades page, star names

The other two pages will follow soon.


In the next posting I will offer discussion and statistical analysis of the 64 star names identified here.


  1. J’ai donné l’adresse dans mon commentaire : hhps://readingvoynich.wordpress.com

  2. Hello everybody ! I think I finally transcribed the three names of the drawing of the Pleiades. See details on my blog please.

    • Stephen Bax

      Can you help people by giving the full address of the page? Thanks

  3. Amanda Budnick

    I was looking at the page with the nymphs in the tub of green iquid and it struck me that it could be the seven sisters constellation-but haven’t of course seen the page in context of what section it is in. So much here to read sorry if its redundant as a comment or already considered

  4. Simon Wayman

    Voynich Manuscript partial analysis of star page 68R

    This is a short summary of my analysis.

    The full Analysis is posted on YouTube as a video;
    Voynich Manuscript partial analysis of star page 68R

    68R3 numbered chart;

    I’ve been using a process of analysing the patterns of star names and position. I have build up this information by using the patterns to cross-link information, names and meaning.
    I’ve been working with Stephen Bax’s ‘sign to sound’ pronunciation, which works very well on this page, and building on his work.

    I have found star names but also a great deal more has emerged, including the variation of spelling in the script, some of the meaning of the charts, and even hints to the origin of the manuscript.

    Star names; (See attached chart)

    Voynich name Modern name

    Tishtrya Aldebaran
    Tir Sirius
    Althoraya Pleiades
    Apaosha Procyon
    Alkhone Alcyone
    Alnt Elnath
    Afyoar Polaris

    Also 2 matching pairs of stars; 5 & 59, 6 & 62.

    There are 2 names, Tishtrya and Apaosha from the Avesta, and Tir from mid-Persian and Iranian language. Also Tishtrya as a Persian Royal star.

    The evidence built towards the charts signifying Taurus as the start of the astrological year, and Pleiades as the Spring equinox. Elements in the charts link to a much earlier date.

    • MarcoP

      I have only now watched Simon’s video. I am happy to see more research based on Stephen’s analysis of the star labels. In this specific case, I am puzzled by the fact that some stars/asterisms are thought to repeat in each of the two illustrations f68r1/2. For instance, Althoraya appears twice in f69r1 and once in f68r2. So the two diagrams cannot be the two halves of a planisphere. What are they?

      The identification of Tystar / Tištrya with Aldebaran mentioned by Simon was proposed in Jean Sylvain Bailly’s Histoire de l’Astronomie Ancienne, 1775 (p.481). This hypothesis was apparently discarded by modern scholars, who identify the star as Sirius.

      I hope that the Voynich ms conference discussed here will actually take place, and that this will be the occasion for an academic expert on the history of astronomy to examine the diagrams and the labels.

    • Neticis

      N at the end probably shouldn’t be ignored too frequently. For example, Russian has quite many words which finish with n, particularly …tan (…тан). Also variations at the ends of words (suffixes) could be inflexions.

    • H

      I think you are onto something with respect to the March Equinox. In another post, I’ve shown the voynichese word on page f57v that I think corresponds to the March Equinox or vernal equinox. But, it doesn’t match the word(s) you think might correspond to March Equinox.

      The term on page f57v appears to be a combination of parts of the words for winter (“ara”) and spring (“dar”). So, it’s possible the language didn’t have a word for the March Equinox. It is also worth noting that EVA “dar” is similar to “ver,” which is Latin for spring. On page f68r3, the March Equinox appears to be implied. I think this page might be showing a transition from (a second word for) “winter” to Taurus.

      So, I think the word you’ve identified might be borrowed from another language (which is consistent with your phonetic method of decoding it).

      I’ve tried to illustrate this below. I also wanted to mention how much I enjoyed your video. An impressive amount of work!

      • MarcoP

        I had previously missed H’s post, highlighting a passage in Simon’s video (37:32).

        The passage discusses EVA:chkchykoly at the top of f68r1. The idea that the Sun and Moon in the “star maps” could correspond to the equinoxes has been discussed several times on this blog (e.g. here).

        Simon and H note that EVA:chkchykoly could read something like yky(u)nkaš(u)n which compares rather well with the Medieval Latin “equinoctium”. Very interesting, in my opinion.

        • Stephen Bax

          Thanks Marco – I agree. Definitely worth considering, as it could corroborate some of our ideas for letter-sound correspondences.

          Of course that word could be borrowed into almost any language.

          • MarcoP

            Hi Stephen, I am glad you find Simon’s observation interesting!
            As always with the VMS, also this idea has its difficulties 🙂
            It seems to me that the match would be better if the word were EVA:chkchyokly / yky(u)nakš(u)n (with the EVA:k and o in inverted positions). Some simple explanations I can think of:

            * The cognate is not Latin but a similar language in which the word for night corresponds better to EVA:ykoly or EVA:koly but the prefix is still similar to the Latin equi- (an unlikely possibility, I guess)
            * The sequence EVA:yk reads differently from what we think (e.g. it simply reads ‘n’ ykynaš[u]n, giving something similar to the Spanish “equinoccio”, or the [u] sound should be inserted differently: ykyn[u]kaš[u]n)
            * The similarity is just a coincidence (this always is an option)

            The hypothetical meaning of EVA:chkchykoly would be validated if we could find a matching interpretation for the preceding label(s) EVA:ykydary.

            • H

              I am still thinking EVA:dary might be some form of “spring.” I noticed that EVA:dary is also a plant label on f88r (thanks to Marco’s link above). So, I started looking for a word that means “spring” that is also a plant and fits within the letter-sound correspondences (which I think is “TURN”).

              I found a map with words for “spring” in various European languages. It lists the word “karun” for spring in Armenian and Google translates spring as “garun” in Armenian.

              I also located a book that (on page 21) lists “Karun” as a type of spring wheat. Although, I’m not sure the plant pictured on f88r looks like wheat.

              I also found a book that identifies Coix lacryma-jobi (Job’s tears) as “garun” in Hindu, which seems like a better fit. Job’s tears is a type of grain native to southeast Asia.

              This source attributes “garun” as referring to Coix lacryma-jobi to Bundelkhand, a region in central India. This book explains that there are a number of different forms that might explain a similar label (EVA:otydary) for a plant on f88v.

              According to this site, the roots and stems each have medical uses.

  5. Val

    Dear Mr. Bax and Mr. Vogt,
    I am also working on the manuscript.
    I think the author wouldn’t want to be read by everyone. when you find something, you don’t know how it can be used. Then I’m still hesitating to show my work. maybe it could be better to go further before taking the decision to reveal It to the entire world. What is you opinion?

    • Val

      I am not talking about linguists of course…

      • Val

        I don’t want to be disrespectful and my english is certainly maladroit.
        Of course I’m not talking about linguistic use. But I think there could be repercutions when this book will be translated. Of course it depends on what is written but isn’t it an important reponsability?
        Maybe I racking my brain (= thinking too much?). But doesn’t it matter? I mean it’s like giving keys of a door to a kid but no-one knows what’s behind…

  6. Peter

    Question? What do Greek gods to do with stars?

    I’ll be at the stars again on Greek names. this can be the connecting something?


    Greek mythology
    1) son of Panopeus, a Greek hero who triumphed over Euryalus at Troy in honor of Patroclus organized by Achilles funeral games in fistfight. He claims to be the strongest pugilist, but admits also that he ermangle of battle fame, because nobody can be all at once. It was he who built the Trojan horse, and even with stepped in.

    2) Epeus, son Endymion, king of Elis, to which the inhabitants of this land called themselves Epeer; his mother was said chromia.



  7. Val

    Dear Mr Bax,
    Mr Vogt found that the eva “t” the letter looking like the latin “M” was a “g” but you are using it like a “l”. I don’t understand.

    • Val

      Is it because of “al atash”? But “al” and “atash” are from two different languages, right?

    • Derek Vogt

      Those are unrelated alternative theories, which were arrived at by different methods.

      • Stephen Bax

        Val, I’m not too sure of what you mean. Which EVA letter are you referring to?

        • Derek Vogt

          Sometimes I hate Arial font. 😀 I presume you must have read “l” as “I” there.

          Reading Val’s “l” as a lowercase “L” instead of a capital “i” makes it clear that (s)he’s talking about EVA-T, which has had both the sounds /g/ and /L/ proposed for it. This is why every time I write either letter alone I use capital “L” and lowercase “i”, so neither looks like it could possibly be the other, regardless of which case I’m using for the rest of the alphabet.

          • Val

            Sorry for the missunderstanding, yes, I was talking about the EVA-T and yes, Arial font doesn’t make it easy…

  8. Hello Stephen, congratulations on your work. I’ve made some visualizations using word vectors, take a look if it this technique is somehow useful for you:

    • Stephen Bax

      Thanks very much. I found your blog posting very stimulating.

      Just to clarify, can you explain what you see as the importance of the fact that the words ‘qoekaiin’ and ‘octhey’ occur both as star names, and also next to each other in folio 48?

      Also, what does it tell you when you note “how close are the star names in the 2D embedding space visualization”? What does that tell us exactly about the VM, in your view?

    • Hello Christian,

      your article is very interesting, but I am a bit suspicious about the graph generated by t-SNE. The two clouds are both almost exactly spherical (although one never knows after the transformation).
      The A/B language classification is not so specific that it would create two clouds.

      I wonder how exactly you entered the transcribed text, because the Voynich MS text has hundreds of free-standing words (labels etc.) while the algorithm works on ‘running text’. It is especially this set of single words that you are interested in. How were they entered?

      However, I am not aware of anyone having used this algorithm before, and the results is likely to be of great interest.

      • Just reading the comments on your blog: I have a tool (C source code) that does all the nasty cleaning up of the transcription file for you.
        Please feel free to E-mail me.

  9. Victorien

    Is it possible that many star symbols are just labeled “s t a l a r” near those they did not know of a name, and a similar word for “starlet”?
    (little “o” equals “s”)
    (Greek looking “PI” equals “t”)
    (Greek looking little “alpha” equals “a”)
    (hand-written loopy “l” equals “l”)
    (Greek looking little “alpha” equals “a”)
    (hand-written little “r” equals “r”)

    I found it interesting the page numbering uses the system 1 2 3 … Arabic numerals.

    • Stephen Bax

      Thanks – but note that the page numbers were undoubtedly added later!

  10. Simon Wayman

    68 r1, r2 & r3 analysis – Summary

    This is a summary of all my previous posts from May 3, 2015 of my findings from these 3 pages. I recommend you read these for all the details and explanations, then this post should make sense.

    Research of ancient star names, especially Tishtrya, tistar, Tištrya is accompanied by the mythology of Pleiades, Aldebaran and Sirius as gods following and fighting in the starry sky.
    The cult of the bull, and bull worship featured strongly for thousands of years and would likely be a big part of the astrology.

    The heliacal rising of stars is the same date each year and was important in the past. This was well known and can be found in many ancient astronomical writings.

    Pleiades’ heliacal rising marks mid-summer. Aldebaran follows Pleiades. Sirius’ heliacal rising marks the start of the rainy season, and the myth is of Sirius, the bringer of rains, fighting its opposite, Apoasha, Apos or Apaos the demon of drought.
    There is confusion though, as historically, Aldebaran and Sirius are given the same name so it’s hard to know what we are looking for on the Voynich charts. The author would, of course, have know the names used at that time.

    The star opposite Pleiades is Antares which has its heliacal rising at mid-winter. This star could be Apoasha as it is important in the myth. Even if it isn’t Antares, Apoasha would seem a good candidate for a Voynich star name.

    Other Voynich charts can be seen to follow the astrological traditions, and it could be expected for the star charts to focus on stars important to this purpose. The chart 68 r3 fits this well for some of these stars, and is the key to the charts.

    The Analysis

    I started without any preconceptions to see anything which could make sense of the charts, apart from Taurus and Aldebaran that Stephen Bax had identified. I let the charts speak to me.
    I began by finding the missing letters of star 59, and a match for it in the text, and then found the exact name again, star 5. Slowly I found more and more close matches of stars, and then these names in the text of 68 r1 & 2 and in the chart 68 r3. Some stars are named 2 or 3 times in the 3 charts as well as in the text.
    Stephen Bax allows for the non-standardisation of language and expects variation in the spelling so I allowed for a small variation and waited to see if any coincidences occurred.

    It became obvious that some of these names appeared with other frequent names so I worked from there to build a picture. From here I have provisionally identified some stars.
    I have marked frequent stars on charts to make it easy to see the matches. (see below & in the next posts)

    The large number of coincidences of these names is far too high to be chance.

    The most frequent star name is 3, 28 & 47 which is linked to the Sun, to Aldebaran, and to star 5, 22 & 59. It’s in the text of the 68 r3 chart 4 times, and without ‘AL’ 3 times. It’s also next to other frequent star names on the other charts and in the texts. Pleiades is the only unnamed object and is linked by a string to this star in the Taurus segment of the 68 r3 chart.

    I think this is very likely to be Pleiades.

    Star 47, a candidate for Pleiades would be pronounced AL YOA TOR

    Swap it round and it becomes AL TOR YOA Al Thurayya?

    I wondered why Pleiades should be the most important object in the 68 r1, 2 & 3 charts. Then while looking through lists of moon mansions I realised that in earlier lists Al Thurayya was the first mansion and so started the astrological year, followed by Aldebaran. (see below) Taurus would then be the beginning of the chart assuming that this is the case.

    Pleiades was often known as Alcyone the brightest Pleiades star.

    Yašts 8 Arabic Moon Mansions
    1. Al Thurayya eta Taurus [Alcyone] The Many Little Ones
    2. Al Dabaran. alpha Taurus [Aldebaran] The Follower

    Hindu Lunar Mansions (nakshatras)
    1. Krittika. eta Taurus [ALCYONE]
    2. Rohini. alpha Taurus [ALDEBARAN]

    Chinese Lunar Mansions (sieu )
    1. Mao. eta Taurus [ALCYONE] Ruled by the Sun
    2. Pi. alpha [ALDEBARAN]

    The sun picture at the top of 68 r1 has a long word on it. The first half of the word is the first word on the page, which might mean sun. It also follows star 3, 28 & 47 Pleiades? in the centre of 68 r3. The next word along is the rest of the word from the sun picture.

    This connection between Pleiades and the sun picture could represent sun in Taurus which would be the spring equinox. It also connects the beginning of the 68 r3 chart with the top of 68 r1. This is a logical way of presenting the charts.

    There is a string of star names down the right side of 68 r3 chart. (see chart below)

    There are star names in the ‘spokes’ of the 68 r3 chart.

    The other stars on the charts which appear important are Aldebaran, star 27, 32 & 48 Sirius?, Star 5, 22 & 59. It can easily be seen these 2 stars and star 3, 28 & 47 Pleiades? are the most frequently mentioned names.

    There are many stars listed in the star charts of 68 r1 & r2 some are listed twice on the same page which means these charts aren’t normal charts of the sky. Apart from some star groupings (circled) there is no apparent logical sequence. They match to the 68 r3 stars a little.

    Also there are a number of stars which don’t match to anything. They are named so must have some importance. The central star on 68 r1 is one of these and could be Polaris by its position. Sirius was moved into Cancer in some astrology charts (due to its heliacal rising) though it’s not on the ecliptic.

    Apart from Aldebaran, Pleiades and Sirius, there are 12 unidentified pairs of stars and 26 single stars on 68 r1 & r2, so some of these must be from outside the moon mansions. So at least one of those charts must contain stars from non-astrological constellations.

    On 68 r2 there is a circle of unnamed stars around the outside, and random unnamed stars in the chart – this must have a meaning.

    There isn’t much text on these 3 pages, excluding star names, so there can only be a limited amount of information given.

    The meaning of the top sun image on 68 r1 is possibly spring or spring equinox.

    The meaning of the moon images on 68 r1 & 2 still isn’t clear.

    The moon in the centre of 68 r3 is likely to mean stars of the moon mansions.

    The 68 r3 chart is in 4 sections possibly representing 4 seasons and or astrology signs. Each segment is 1 ½ astrological signs wide so it doesn’t just refer to the signs.

    An interesting observation
    Star 21 has an unusual flourish on one of its letters. The matching star 30, also has this. Look carefully at the second flourish and you can see that the scribe has stopped, probably to check the first one and then continued, finishing it in exactly the same place. Presumably to make the names look similar but there are some differences. This could show that they allowed quite a lot of variation or that the scribe might have done this deliberately to show the two names are the same, as they don’t appear elsewhere on the charts.

    I have produced some charts of this information showing what I think to be reasonable matches of star names. Its now up to you all to see if you agree with them and can make sense of them. At least I’ve made a start for you to explore. There has to be some correspondence between these names on the charts or the charts will be pointless.

    There are many more possible connections I haven’t mentioned. In fact I’m seeing new things all the time. There’s more to extract from these charts and they are giving up their secrets bit by bit. I feel that rather than trying to work out the possible meaning, its best to look and see what they have to say. I have found connections that I never expected.
    I am relying on help from other Voynich scholars, and especially Stephen Bax to provide help.

    Below is a chart 68 r3 with star matches. Charts for 68 r1 & 2 and also my chart of numbered words in 68 r3 will be in the next posts.

    For your interest;

    I didn’t use my computer very much to analyse these pages. I looked very carefully at the downloaded charts. I found that by printing out each page on an A4 sheet they are much easier to read than on a screen, and easy to compare names by moving the sheets together and rotating them where necessary, also for writing notes.

    You can download the pages in hi res from the Beinecke web site.

    Go to Beinecke site and you can download the whole manuscript as a PDF which is mostly of high quality. The larger fold-out pages are of lower resolution but you can download individual pages at hi res one at a time – very worthwhile. The ordinary pages are already of this resolution.

    Beinecke Library Rare Book and Manuscript
    Low down on the right of the page click on ‘Export as PDF’ to download the whole manuscript.

    From the same page scroll through the lower pictures of the manuscript to the page you want. Click on the image and it will appear above. Click on ‘zoom view’ and the hi res image will appear, click on ‘download image’ and save it as a jpg.

  11. Simon Wayman

    Matching stars 68 r2

  12. Simon Wayman

    Matching stars 68 r1

  13. Simon Wayman

    Star matches 68 r1

  14. Simon Wayman

    Chart 68 r3 numbered words

  15. Simon Wayman

    Sun & Moon 68 r1 & 2

    68 r1 shows a bright Sun with orange flames at the top which might indicate sun rise, summer or mid-summer.
    68 r2 shown a dark Sun at the bottom with blue flames which might indicate sunset, winter or mid-winter.
    The meaning of the Moons is not clear.

    The first word in the text of page 68 r1 can be found as the first part of the word on the bright Sun drawing (same page). This might be the word for Sun.
    The same word can be found on 68 r3; 48 (my chart). The next word along is the rest of the word on the Sun drawing.
    The word preceding 68 r3;48 is star 3, 28, 47 which is linked by the string to Pleiades. Yet another interesting connection to this star.

    68 r1 under the Sun drawing are star 3, 28, 47 Pleiades?, and star 5, 59 Antares?, both have an important heliacal rising – mid-summer and mid-winter.

  16. D.N. O'Donovan

    For people interested in the star-names around the month-roundels I would suggest reading Varisco’s works on the anwa’.

    • Stephen Bax

      This website has a good discussion of the “anwa'”:


      • D.N. O'Donovan

        Thank you for taking that trouble. Serjeant is actually the person who opened up this field of study, but more recently Varisco has produced a lot of papers and books, so his is easier to find. I prefer his more recent works to his earlier essays.

  17. Simon Wayman

    More stars

    New star pairs;

    Star 6 & 62
    Star 7 & 20
    Star 21 & 30

    4 stars 10, 26, 38 & 56

    Star 22 added to star pair 5 & 59

    Matches in text;

    Star 1 & 24 68 r3;A5
    Star 35 68 r3; E3

    The Pleiades heliacal rising marks the summer solstice so we could assume the star opposite it on chart 68 r3 might mark the winter solstice. Star 5, 22 & 59 is perfectly placed for this.
    These stars appear together, star 3, 28, 47 (Pleiades?) 68 r3; A1 (my chart) is next to star 5, 22 & 59, 68 r3; A2.
    3, 28, 47 (Pleiades?) also appears at the top of the chart 68 r3;1.

    The modern astronomical name for the star which has its heliacal rising at the winter solstice is Antares, so this could be star 5, 22 & 59.

    • MarcoP

      The modern astronomical name for the star which has its heliacal rising at the winter solstice is Antares, so this could be star 5, 22 & 59

      If f68r1 is a star map, how can Antares be star 5 and 22 at the same time? Anyway, #5 as Antares would be consistent with Stephen’s identification of #29 with Tashyata / Taschter/ Tishtar / Sirius. If we read f68r1 as a map of the Southern Sky, we can also match Star #24 (EVA:okoldy, reading something like “Akashtun”) with Canopus, whose Hindu name is “Agastya”. Star #27 (EVA:ordaiin Artaur?) could then correspond with Adhara. The central star #16 could match the Southern Cross, the South Pole of Hindu astronomers).

      It would be interesting to develop an alternative to the extensive analysis detailed by Derek, but this new hypothesis is currently quite weak:
      * The three labels for stars #29, 27 and 24 are based on three different languages (Persian, Arabic and Hindu)
      * The positions of Antares and the Southern Cross could fit, but I could not find any possible interpretation for the labels.

      • Johannes Klein

        Hi All,
        is it not much more likely that f68r1 and r2 represent two different areas of the same sky seen at one location, than the N and S hemisphere. As for example the winter and summer sky.

        • MarcoP

          Hello Johannes,
          couples of circular diagrams representing the Northern and Southern hemispheres of the “sphere of the fixed stars” appear in numerous medieval manuscripts. As far as I know, all these maps include images of the constellations, which are not present in f68r1/2. Still I think those manuscripts are good parallels for the two Voynich images.

          An Islamic tradition of celestial globes representing only stars (without constellation images) is also well documented (Savage-Smith lists exemplars dating from the XII to the XIX century). I think that those globes also are interesting parallels for the f68r1/2. Of course, celestial globes and couples of circular images as those linked above are just different ways or representing the same thing (a complete map of the sky).

          A possible analogue to your proposed interpretation could be provided by the tympans of astrolabes, which represent heavenly coordinates corresponding to different observing latitudes. But the tympanes do not include the positions of the stars (only coordinates grids). In an astrolabe, the layer representing the fixed stars (the “rete”) usually is only one (while in the Voynich ms we have two similar “star maps”). Astrolabes having two retes (Northern and Southern fixed stars) are again analogues of the manuscript illustrations mentioned above: a couple of items representing the two halves of the celestial sphere.

          These parallels (in particular the manuscripts and the stars-only celestial globes) suggest to me that the extension of Stephen’s analysis proposed by Derek and others is a very sensible approach.

          Could you please explain why you think that f68r1 and f68r2 represent “two different areas of the same sky seen at one location”? Why is this interpretation more likely than the others? Have you ever seen anything like that in ancient manuscripts? If the two diagrams represent the Winter and Summer sky seen from the same location, shouldn’t they have a significant central overlap?

          • Johannes Klein

            Dear Marco,
            I guess I was mislead by your previously attached picture that you were taking the two circular diagrams as representing true N and S celestial hemispheres. In the sense of: a projection of the nightsky with the respective pole in the middle delimited by the equator, so that the diagram exclusively shows the N or S hemisphere respectively (which correspond to the sky seen at earths north pole and south pole, respectively). From the MSs posted in your reply I see that this is not the case. However, these represent two different traditions: link 1 and 2 is similar to what I mean with summer and winter sky, in the sense that one diagram focusses on the constellations that are visible in the winter sky from a given locality on earth (i.e., at discrete latitude) and one on those in the summer. I am not an expert in astronomical maps, but I was guessing that this kind of drawing was more common in the time of origin of the VMS, and that the kind of representation as seen in links 3 and 4 appeared later. My Italian is not so good anymore, but does the website say if this type of drawing was rather common for that time?

            • Johannes Klein

              UPenn LJS 57 (link 3 of Marco’s earlier post)
              Also has a small star section on pages 1 and 2.


            • MarcoP

              Hello Johannes, I think that couples of planispheric maps of the heavens were rather common in astronomical manuscript. You can read about the subject in Celestial Mapping, an essay by Emilie Savage-Smith. Even if the essay focuses on the Islamic tradition, at p.60 the author discusses the 1440 Vienna manuscript I linked above.

              At p.13 she writes that, as early as the VIII Century, planispheric maps of the heavens could be found in a number of Latin and Byzantine manuscripts. She also writes that “Despite the large number of medieval Islamic manuscripts preserved today, it is notable that none of them contain planispheric maps of the sky”.

        • Derek Vogt

          I don’t think anybody picked one model or the other to begin with. I know I didn’t. At first, we were just looking for known star names that would be similar to Voynich star names, regardless of where they were on the pages. Once there were a few good candidates, identifications became easier for other nearby stars based on proximity to the ones that had already been identified before, so star identifications gradually spread across the map from neighbor to neighbor, one at a time. The model of celestial hemispheres, instead of the model of seasonal maps from a single viewpoint, is simply what emerged from that process.

      • MarcoP

        On the basis of the recent growing interest in Romany and Indian languages in general, I will slightly expand what I wrote one year ago.

        voynichese.com suggests that EVA:otochedy only occurs as the label of Star #28 and in the central text circle surrounding the Cancer zodiac sign (f72r3). The word appears immediately after the symbol marking the beginning of this text fragment. One could speculate that the label means “Cancer” (or denotes a prominent star making part of the Cancer constellation).
        Stephen reads the label as “alayot(un)”, possibly corresponding to Al-Yad (name of stars in the Ophiucus and Orion constellations).
        Derek Vogt reads the label ^agahotn^ and interprets it as a reference to Scorpio. Interestingly, he observes that in some languages the words for “scorpion” and “crab” can get mixed up. He also mentions the Nepali word for “crab”, गंगटा “gaŋgaʈā”. So it seems that Derek independently noticed (on a purely linguistic basis) a possible connection between Star #28 and Cancer.
        I think that also the Sanskrit name of the sign of Cancer could be a relevant cognate: Karkata. Also the Tamil: Karkatan (mentioned by Allen). I roughly mark the position of the Cancer constellation in orange (it is in the Northern Sky, just outside the scope of this map).

        In this page, Stephen proposed that the label of Star #29 (EVA:dolchedy) could read something like “Tashyata”, corresponding to the Indo-Iranian Tishtrya / Tishtim (Vedic tiṣya) identified with Aldebaran or Sirius.
        As can be seen in the attachment, Stars #28 and #29 are close to each other, and also the Cancer constellation and Sirius are relatively close.
        In Hindu astronomy the term Tishya also identifies an asterism included in the Karkata constellation: this represents a third alternative to Aldebaran and Sirius.

        Star #24 (label EVA:okoldy). The label is read similarly by Stephen [akasht(u)n] and Derek ^ãkaštn^. Derek interprets the label as a reference to the Greek-Latin Arktouros/Arcturus in the Bootes constellation. As I wrote in the previous comment, a different possible interpretation would be the Vedic sage Agastya, also identified with the star Canopus the brightest star in the southern constellation of Carina, and the second brightest star in the night-time sky, after Sirius.

        • Derek Vogt

          So the idea is to start over and use these as the basis for an entirely different set of identifications for the whole page, using the left side as the southern celestial hemisphere?

          Or is it that the names of a line of southern stars extending south from near Cancer got transferred to some northern stars in a line extending north from near Cancer?

          • MarcoP

            Hello Derek,
            the idea is just sharing facts I find interesting, hoping that someone will find them useful. One year ago I wanted to mention Agastya / Canopus. Currently, I find the connection between Cancer and Star #28 more interesting. Obviously, since Cancer is on the ecliptic, this hypothetical identification does not suggest a specific North / South interpretation for the map. Sirius could conceivably be represented in the Northern map, or the label of Star #29 could correspond Aldebaran or to the Hindu asterism making part of Karkata, then properly belonging to the Northern sky. I tried to arrange this facts into a different reading of the map, but I am not at all confident that this idea makes sense.

            I think that, if we could detect some internal coherence between the star maps and the zodiac pages, the two sets of data could maybe support each other’s interpretation.

            I am ignorant of Vedic astronomy, but since other linguistic considerations have highlighted the relevance of Indian languages, I also think that this area deserves further investigation. This is a subject I need to study more deeply.
            For instance, after posting my comment I read on wikipedia that the Vedic sage Agastya is also identified with one of the stars of the constellation Saptarshi (Ursa Major). So also this specific reading of Star #24 does not necessarily imply that the map represents the Southern sky: all these things are new to me.
            I also found out that, according to this source, in Indian astrology only Scorpio and Cancer are or reptile “nature” (or “yoni”): this could possibly be a sufficient explanation for the occurrence of EVA:otochedy only in the Cancer diagram and near the star you identify as Scorpio, even if I don’t remember ever seeing a constellation only labeled by its nature. And what is the Hindi word used for “reptile” in this context? Being familiar with Latin astrology, I can see that “casta” (and even more clearly “castum”) cannot be used in place of “virgo”, even if it has a similar meaning. I am currently unable to say anything about the possibility that ^agahotn^ corresponds to the “reptile” Yoni of Indian astrology.

  18. D.N. O'Donovan

    For your interest,
    A fourteenth century star-map (planisphere) from somewhere else. It would be fun to read these star-names Romanised, wouldn’t it.

  19. Simon Wayman


    I’ve found more of star 3 & 28 (Pleiades?) in the text of 68 r3 this time without the ‘AL’ (See my chart) 68 r3; 22, 36 & 37.

    New match for star 3 & 28 (Pleiades?) is star 47. This version is spelt with ‘TAR’ on the end.

    Looking at the name for this star now it would be pronounced AL AYO or YOA TOR

    Swap it round and it becomes AL TOR AYO or YOA
    This is a good match for Al Thurayya.

    This feels like a composite of 3 words; something like ‘The Taurus Group’ or ‘The Taurus Stars’. This comes from matching patterns not from linguistics.

    New match for star 27 & 32 (Sirius?) is 48. This is 2 words, the 2nd is a near match for the last part of the ‘string’ word.

    This star now forms a group on 68 r2; star 43, 29 & 54 Aldebaran, star 46 (68 r3; 7 & E1), star 47, 3 & 28 (Pleiades?) and star 48, 27 & 32 (Sirius?).

    We now have 2 groups with the 3 most common stars, one on 68 r1, one on 68 r2.

    More to follow

    • MarcoP

      Hello Simon, thank you for sharing your research about the possible matches linking the three f68r diagrams! I tried a simple match between EVA labels for f68r3 and the other two diagrams (using Levenstein distance, not necessarily linguistically significant). These are the matches I find more sound (distance 1 and 2). They include your hypothesis that star #59 matches star #5, as well as the possible identity of star #54 and #43 discussed by Stephen in his April 2014 talk (please note that Tishtar is usually identified with Sirius, not with Aldebaran). I think you have already mentioned most or all these possible matches:

      r3:54 dcholday -> dcheoldy (r2:43) d:2
      r3:57 otory -> otor (r1:19) d:1
      r3:59 chocphy -> chocfhy (r1:5) d:1
      r3:60 okolchy -> okoldy (r1:24) d:2
      r3:61 otydm -> otydy (r1:9) d:1
      r3:64 okchody -> okcheody (r2:41) d:1

      I hoped to find some clear visual pattern in the angular positions of the matching stars, but at first sight I cannot make sense of the matches. Being able to find some kind of internal coherence between the labels of the three diagrams would be an important step forward!

      • Johannes Klein

        Dear Marco,
        without being to pedantic, I would argue that the label, with three exceptions (#25, #36 & #63) always is to the right hand side of the star in question. So that your coloring for star #43 should be one down to the left.

        • MarcoP

          Thank you, Johannes!

          • Johannes Klein

            Dear All,
            prompted by Marco’s comment, I was looking for a pattern in the number of tips of the stars on the f68r diagrams. I colored the stars according to their number of tips in the below figure. My conclusions are the following:
            1. There are stars with 6-9 tips. 68r3 only has 6 or 7 tipped stars.
            2. All labeled single stars in 68r3 have seven tips.

            1. The four parts of 68r3 that are densely filled with stars have 6 and 7 tipped. Number of tips in these parts seems not to be correlated with available space to draw.
            2. In 68r1 and 68r2 stars with the same number of tips are not clustered. My Interpretation for these two diagrams: Position of a given star within the diagram in combination with their number of tips is either completely random or completely non-random. For the fact that there are no 8 and 9 tipped stars in 68r3 (the other diagrams with stars remain to be checked) I would argue for the latter case.

            My overall conclusion is: For 68r3, the position of a star within(!) one of the eight segments is random. Whereas in 68r1 and r2 the position is non-random. Whereas, therefore, 68r3 may be a thematic grouping, I would argue that 68r1 and 68r2 represent actual maps of stars, i.e., their relative positions have some counterpart in the natural world (in the sky).

            Hypothesis: The number of tips on a star drawn in 68r1 and 68r2 represent some form of category of stars. What comes to my mind at first would be brightness/magnitude.
            Remains to be tested.

            If 68r1 represents an actual map of the sky on additional observation I would like to point out (without knowing if this has been written somewhere before):
            stars 8-9, 13-14-15, 23, 22-24 together bear a remarkable similarity to the brightest stars in the constellation Orion. In this case, however, the brightness would not be depicted as the number of tips, as 24 (Rigel) would be the brightest star.

            • Johannes Klein


  20. D.N. O'Donovan

    I hope I’m not duplicating an earlier comment – the list is rather long, and my time today rather short.

    The stars in the margin are, I think, a way of rendering the same convention that is common in Islamic manuscripts in Persian or Arabic, but chiefly the former, by which the star-flower divides paragraphs or sententiae. The Persian style doesn’t draw the stars as they are in MS Beinecke 408, which represents the six-pointed “syrian” star, as distinct from the invariably 5-pointed star of ancient Egyptian art, or the 8-point of Babylonia and so forth.

    The Persian ‘asterisk’ is much more rounded and usually has eight or more ‘petals’. There are examples everywhere, but my ‘theriaca’ posts include some nice examples.

    The stars-on-strings around the month-roundels are probably astrolabe (‘star-holder’/star-gripper’/’star-taker’) stars, though I shouldn’t bank on their being named by the so-called ‘Arab’ star-names, or by the older Latin descriptions.

    Still, the language isn’t my area.


  21. Simon Wayman

    More on Taurus

    Looking at 68 r3 Taurus segment and the centre we see Taurus 68 r3; 50 (see chart in my May 24th post) and a word under the string from the Pleiades 68 r3; 49.
    If you follow the string it points to 68 r3; 47 which I have identified as star 3 & 28. (Appears twice at the top of the page)
    From 68 r3; 47 move clockwise and the 3rd word along is the 2nd half of the string word 68 r3; 49. Next to that is a short form of Taurus 68 r3; 41. (see picture below)

    Pleiades is not labelled like all the other stars but is linked by a string to a name which is a star.

    Possible identification of star 3 & 28 is Pleiades. However, its only a vague match to the ancient names we have for it.

    In 68 r1 Stars 27 & 32 (Sirius?), 3 & 28 (Pleiades?) and 29, 43 & 54 Aldebaran are all together.

    On the right side of 68 r3 chart we have a cascade of star names linking stars 62, 46, possibly Aldebaran, Sirius, Pleiades, TAR + most of Pleiades, and Pleiades. (see picture)

    How could we have missed it!

    I think that its plausible as there are so many cross links. Of course it will have to be scrutinised and changes made where necessary.

  22. Simon Wayman

    More star matches

    I’ve been working through 68 r1, 2 & 3 carefully and have found matching pairs of stars in the charts of ‘named stars’ and also matches in the text above r1 & r2, and in the r3 chart. These names are almost all very near matches with a few exact ones. I’ve only included the closest matches (perhaps one letter difference) to give the best chance of identifying the stars.

    Matching stars; (L= line W= word)

    Star 3 & 28 Text 68 r2; L1 W8. 68 r3; A1, 1, 10, 12, 47 (see chart from my last post)

    Star 41 & 64 (Almost same as star 3 & 28)

    Star 18, 19 & 57 Text 68 r1; L1 W7

    Star 9 & 61

    Star 10, 26 & 38

    Star Pairs; all on 68 r1

    Star 1 & 24

    Star 3 & 28 (see above)

    Star 5, 22 & 59

    Star 18, 19 & 57 (see above)

    Single stars also in the text;

    Star 46 Text 68 r3; 7, E1

    Star 62 Text R3; 6

    Star no.16 might be Polaris by its position. If you look carefully at the drawn star next to the name, there is a small dot in the centre. I found that it’s the hole left by the pair of compasses used to draw the outer circle. So the star has been drawn ‘exactly’ in the centre. It could also verify 68 r1 chart as being the northern sky.
    Its only shown once on the chart and once in the text 68 r1; L1 W6 (possible match).

    More to follow

  23. Simon Wayman

    68r 1, 2 & 3

    With reference to Stephen Bax’s numbered star charts.

    From my last post there are now 2 stars which appear twice on the charts, and a star name in the text of the charts.

    My candidate is star no.5 & 59, the other is Aldebaran star no.43 & 54 which Stephen identified.
    He allowed for the non-standardisation of language by identifying the two stars as the same even though there is a difference. I have found another possible match for Aldebaran with star no.29 if we allow enough flexibility in the spelling. That means its on all 3 pages but it is an important star.
    Near matches for it can also be found in the Voynich text, 68 r2; line 4 word 8, and 68 r3; F4 (see my chart below).

    (From now on I will use L for line & W for word i.e. 68 r2; L4 W8 as above)

    There are many variation of the name from historic times such as Tishtrya tistar Tištrya Tishtryaeninis Tascheter so perhaps its not surprising it varies in the Voynich.

    The heliacal rising of the Pleiades was important because it would disappear behind the Sun for months, then reappear just before sunrise on the same day each year, and that day marks the Summer Solstice. Aldebaran would follow with its heliacal rising shortly after, hence ‘the follower’. Its interesting that although Taurus is a spring constellation, with the Sun passing through Taurus in April/May, Pleiades rising marks mid-summer.

    Sirius – The brightest star in the sky
    From an Iranian astrological source, Sirius has been confused with Aldebaran and the name Tishtrya used for both at times. Perhaps because Sirius also has a heliacal rising, which is in early August and heralded the coming of the rains. According to this astrology Sirius was and still is, in the constellation of Cancer. This was due to the time of its heliacal rising when the Sun was in Cancer. (In modern astronomy its in Canis Major and called the Dog star)
    The other name for Sirius from this time was Tir or Tiri. As Tishtrya was a match for Aldebaran in the Voynich, then Tir or Tiri should be a match for Sirius.
    I found star nos. 27 & 32 to match if we can omit the ‘o’ and ‘or’ at the beginning. The ‘or’ on star 27 could be a separate word.
    Eva version; or!daiin & odaiin possible pronunciation TAIR which is close.

    I found this word twice in the text of 68 r3 (See my chart) F3 & 9.
    F3 is in Cancer on the chart and is also next to Aldebaran F4.

    I think it likely that the ‘named stars’ on the charts, when their names vary a little, are not different stars. Thinking of Aldebaran, would they have had 3 stars with almost the same name?

    The probable meaning of the charts on pages 68 r 1, 2 & 3 is to help find and identify stars and show their relative position. There should therefore be mention of the stars in the text to assist this. These stars will be important to the author and signify important dates like; solstice, equinox, religious festivals etc. It should be possible to home in on star names which appear most frequently, and to see their significance.

    Also by examining the position of the stars on the charts it could be possible to say something about what the charts are for. This can work even if we can’t identify the Voynich star names. If stars appear on r1 and r3 then perhaps r1 isn’t just the northern sky, or includes lower latitudes, or maybe something else. Patterns will emerge.

    I should say that I’m not using web sites and computer searches much for the matching of names. I’m using my eyes and comparing very carefully. The results might not appear to agree with others’ opinions. For instance, much that I appreciate and use the voynichese site there are many discrepancies compared to what I see in the manuscript.

    More to follow

  24. Simon Wayman

    68 r3 star no.59 missing letters found

    There are some missing letters in this name where there is a fold in the page. I found an earlier high resolution photo of the same page which has less damage. (See top image)
    I have managed to find the name elsewhere on the same page. I’m sure it’s the same. (See bottom image)
    Its the 2nd word down vertically from 12 o’clock on the 68 r3 page.
    What is even more interesting is that the name is an exact match with star no.5 on Stephen Bax’s numbered star pages.

    I expect there to be mention of the star names in the text to explain the charts’ meaning.

    More in next post

  25. Simon Wayman

    Hi Stephen. Thanks for your fascinating work on the Voynich.

    Possible star patterns and meanings for 68r3.

    This is to attempt to make sense of the star pages from an astronomical viewpoint. There is some guesswork and assumption here but it’s using logical steps which, if they prove wrong allow for another logical step. I hope to narrow down the possible star name by this method.
    I’m also looking for patterns, why things are laid out in a particular way. Trying to get into the mind of the author.

    Having read your translation of The Taurus segment of this page I have researched the possible meaning of the chart. As it has very little writing on it, it may simply be a pointer to the stars of the seasons. Taurus for Spring etc for the 4 quadrants with named stars. I have ignored the 4 quadrants full of unnamed stars.
    On modern star maps of the ‘constellation of Taurus’, Aldebran and Pleiades are very close to the plane of the ecliptic, so I guessed that it is aimed at astrological constellations.

    (The ‘plane of the ecliptic’ is the path the sun takes through the sky as seen from the earth. Marked in blue on the star charts. The astrological constellations are all on this line).

    Therefore I chose constellations every 3 months of the year starting with Taurus. Heading anticlockwise on the chart, the next quarters we have Leo; Summer, then Scorpius; Autumn and Aquarius; Winter. (If it goes round clockwise on the chart then Leo and Aquarius reverse).

    The Leo, Scorpius and Aquarius star quadrants have 4, 3, and 2 named stars in a pattern (which may be significant). We can now try to put known stars into those patterns and see if we can match them to known patterns or names. The pattern for Taurus matches the actual pattern on the star map exactly. (see Taurus below)

    In the Voynich Taurus quadrant, only Aldebran and Pleiades are shown. The nearby, and obvious Betelgeuse and the 3 stars of Orion’s belt aren’t included, so perhaps only the brightest stars (or objects) of that constellation, close to the plane of the ecliptic, are shown. I assume the same would be true of the other 3 quadrants. I have listed the most likely stars from each constellation.

    In modern astronomy the stars of a constellation are given names ie in Scorpius; α Scorpii (Alpha Scorpii), β Scorpii (Beta Scorpii) etc. with the brightest named first. They have common names as well sometimes. I have ignored stars without common names.

    I have not looked for planets here because they don’t fall into neat yearly patterns.

    Listed below is information about each constellation, with links to info and ancient names of stars. I used the very useful star charts from http://www.constellation-guide.com. There is a wealth of info here and I quote from it below. Also http://earthsky.org/constellations is a useful site.

    Taurus Spring; April 20 – May 20

    Info & map

    Look at the Map to compare the star pattern.

    2 objects on diagram Good pattern match.

    Leo Summer July 23 – August 22

    Info & map
    More info & photo

    Look at the Map to compare the star pattern.

    4 stars Good pattern match
    Named below are likely stars
    Regulus, α Leonis Alpha Leonis
    Denebola – β Leonis Beta Leonis
    Algieba, Gamma Leonis
    Adhafera, Zeta Leonis

    Gamma Leonis is a double star in Leo. Its traditional name, Algieba or Al Gieba, comes from the Arabic Al-Jabhah, which means “the forehead.” The star is sometimes also known by its Latin name, Juba.

    Scorpius Autumn October 24 – November 22

    Info & map
    More info & photo

    Look at the Map to compare the star pattern.

    3 stars on diagram. Good pattern match.
    Named below are likely stars
    Antares – α Scorpii (Alpha Scorpii) [well known bright star]
    Acrab (Graffias) – β Scorpii (Beta Scorpii)
    Dschubba – δ Scorpii (Delta Scorpii)
    Shaula – λ Scorpii (Lambda Scorpii) [unlikely]

    Antares comes from the ancient Greek Άντάρης, which has been translated as “anti-Ares,” “rival of Mars,” or “like Mars,” referring to the similarity of the star’s red hue to that of the planet Mars. Another theory suggests that the name Antares may have come from the name Antar or Antarah ibn Shaddad, which was the name of an Arabic warrior-hero celebrated in the Golden Mu’allaqat, one of the seven long Arabic pre-Islamic poems.

    In Persia, the star was known as Satevis, one of the “royal stars.” In Mesopotamia, the names associated with the star include Dar Lugal (“the King”), Masu Sar (“the Hero and the King”), Kakkab Bir (“the Vermilion Star”), Urbat, Bilu-sha-ziri (“the Lord of the Seed”), and Kak-shisa (“the Creator of Prosperity”).

    To the Egyptians, Antares represented the scorpio goddess Selkit (Serket), and also served as the symbol of Isis in different ceremonies. The ancient Greek name for the star, Καρδιά Σκορπιού (Kardia Skorpiū), meaning “the heart of the scorpion,” was translated into Arabic as Qalb al-Άqrab, and Arabs knew the star as Calbalakrab. The name was also translated into Latin as Cor Scorpii, one of the names still used for the star.

    Babylonian astronomers listed Antares as GABA GIR.TAB in MUL.APIN. The name means “the breast of the scorpion.”

    Antares forms a group of stars known as the “Royal Stars of Persia” together with Aldebaran in Taurus, Fomalhaut in Piscis Austrinus, and Regulus in Leo constellation.

    Acrab (Graffias) – β Scorpii (Beta Scorpii)

    Dschubba – δ Scorpii (Delta Scorpii) traditional name, Dschubba (or Dzuba) comes from the Arabic jabhat, which means “the forehead,” referring to the scorpion’s forehead. Sometimes the star is also known as Iclarcrau or Iclarkrav.

    Shaula – λ Scorpii (Lambda Scorpii) is a multiple star system with three visible components. Its traditional name, Shaula, comes from the Arabic al-šawlā´, which means “the raised (tail).”

    Aquarius Winter January 20 – February 18

    Info & map
    More info & photo

    Look at the Map to compare the star pattern.

    2 stars on diagram. Good pattern match.
    Named below are likely stars
    Sadalmelik – α Aquarii (Alpha Aquarii)
    Sadalsuud – β Aquarii (Beta Aquarii)

    Sadalmelik – α Aquarii (Alpha Aquarii) Its name is derived from the Arabic phrase sa’d al-malik, which means “luck of the king.” Sometimes the star is also called Rucbah, a name it shares with Delta Cassiopeiae.

    Sadalsuud, β Aquarii (Beta Aquarii) Arabic phrase sa’d al-suud, meaning the “luck of lucks.”

    Looking at the other 2 pages, R68r1 &2, I wondered if the meaning of the sun and moon near the stars could mean stars visible in the evening/early night near the moon, and towards dawn near the sun. On the R68r3 page the moon in the centre might indicate evening constellations, which is what I found to be true.

    In R68r1 &2 the moon and sun have swapped places, possibly denoting summer and winter, and the stars associated with them. The sun at the top might mean sunny days – summer, the moon at the top, dark – winter.

    Interestingly none of the names in R68r1 are in page R68r2, which might aim to show as many star name in a limited space.

    The stars of a well known group may be shown together on pages R68r1 & R68r2, so in Orion we should see Betelgeuse, Bellatrix, the 3 stars Belt stars – Alnitak, Alnilam, Mintaka, then Saiph and Rigel, and nearby Sirius. Bear in mind they may have had different, ancient names.

    I also noticed in R68r1 there seems to be a star at the very centre of the circle. Stephen Bax star no.16. Could this be the star that all others rotate around – Polaris? Very important for navigation.

    Info from;
    Names for Polaris;
    Alruccabah; Cynosura; Phoenice; Lodestar; Pole Star; Tramontana; Angel Stern; Navigatoria; Star of Arcady; Yilduz; Mismar.

    In traditional Indian astronomy, its name in Sanskrit is dhruva tāra “fixed star”. Its name in medieval Islamic astronomy was variously reported as Mismar “needle, nail”, al-kutb al-shamaliyy “the northern axle/spindle”, and al-kaukab al-shamaliyy “north star”. The name Alruccabah or Ruccabah that was reported in 16th century western sources was that of the constellation.

    Modern star names for common bright stars – brightest first

    Sirius Canopus Arcturus Vega Capella Rigel Procyon Betelgeuse Achernar Hadar Altair Acrux Aldebaran Spica Antares Pollux Fomalhaut Deneb Mimosa Regulus Adhara Castor Gacrux Shaula Bellatrix Elnath Miaplacidus Alnilam Alnair Alnitak Regor Alioth Kaus Mirfak Dubhe Wezen Alkaid Sargas Avior Menkalinan Atria Koo She Alhena Peacock Polaris Mirzam Alphard Algieba Hamal Algol Mizar Mintaka

    http://domeofthesky.com/clicks/alphalist.html This site has ancient names and meanings.

    Best wishes, Simon

  26. MarcoP

    This pdf by Günther Oestmann lists the 35 stars that appear on the 1493 celestial globe by  Johannes Stöffler. The names are mostly Arabic. Coordinates are also given for each star.


    • MarcoP

      This paper, also by Günther Oestmann, provides a more complete list of the stars on Stoeffler’s 1493 Konstanz Celestial Globe. 47 stars are named on the globe: rather close to the 53 labeled stars on f68r1/2.

  27. It is an ancient tradition in northern Europe for women to wear large, circular brooches called solde, solju, or risku. These brooches represent the sun. Often, they have smaller suns dangling from the central one. Only women of child-bearing age wear these brooches, because the sun symbol is tied to fertility. This tradition reflects a very ancient belief in a female sky deity who gave fertility to the earth. The myth about a heavenly maiden is known in the mythology of the Sami, Estonians, Mari, Komi-Permyaks, and is reflected in traditions surrounding Frau Holle, Freya, and Hulda. The dísir, who play various roles in Norse texts, are explicitly called dead women in Atlamál 28 and a secondary belief that the disir were the souls of dead women (fylgjur) also underlies the landdísir of Icelandic folklore. They were a sort of protective spirit for their clans. In such a tradition, the dead female ancestor who founded the clan would be depicted as a sun in the center of a star chart. The Voynich manuscript with its heliocentric, female star charts was written 105 years before Copernicus’ publication that posited the sun in the center of our planetary system, causing the Copernican Revolution. More than likely, you’re not looking at anything astronomical at all but rather genealogical. The manuscript is not classical. It is not Islamic. It is not Indian. It is not even Nordic in any comfortably familiar way. But it is firmly grounded in a certain tradition that scholars keep overlooking.

  28. A late note.
    A nineteenth century Indian globe has very little practical purpose, and was most likely made as an ornament – such as for a gentleman’s library.

    If you go back to instruments made between the tenth and fifteenth centuries, you find that the number of stars is always rather small by our standards. I don’t have my notes with me, but some time ago I compiled a list by surveying the texts in conjunction with those astrolabes which are described in Gunther’s 2-volume work. As I recall, the number ran between 17 and 32, with most of the earlier ones having no more than 15, and even those made to the end of the fifteenth century rarely having more than 22.

    “The fifteen stars” which Cornelius Agrippa made so much of in the sixteenth century as ‘Behenian’ stars were the usual list taught long before the sixteenth century.

    A useful site (if you’re interested in astrolabes)

    My own opinion is that the people who made most of the astronomical imagery had never needed to use brazen instruments and had studied their stars more directly than any medieval European before the sixteenth century.

  29. MarcoP

    Many thanks to Stephen for finding and sharing the Arabic T and O diagrams. They could support the hypothesis that Voynich T&O illustrations are not maps but representations of the elements. In particular, I think it is interesting that “Earth” is possibly denoted by similar words in the Arabic diagram and f68v3: the Arabic al turab and the Voynichese turash (EVA:darol). I searched for languages in which “turab” became “turash” / “turas” / “turaz” / “turaf”, but with no success. So I started thinking about word endings in general.

    I would like to know the opinions of other people about this fact: we are considering Persian and Arabic as possibly related to Voynichese. In those languages, words ending in -f, -b and -l are rather common (about 20% of the total words). On the other end, Voynichese words ending in EVA:p / EVA:f / EVA:t are extremely rare (so rare that the few occurrences could all be transcription errors). How can this phenomenon be explained?

    In particular, I would expect that (if all star names derive from Arabic names) about 12 of the 64 star names we are considering derive from originals ending in -f, -b or -l. I think that the only possibilities are:
    1. suffixes (e.g. EVA-y) were appended;
    2. the ending consonant was dropped;
    3. the ending consonant was changed into a different sound (as with “al turab”, in the hypothetical reading of the T and O diagram).

    I have checked the 64 star names for endings in EVA:p/f/k. There are none. Ignoring the last character, I find a single match. Ignoring also the second to last character, I find a total of 6 matches:
    5 chocfhy
    15 ocphy
    22 ockhy
    56 okos
    58 chdy.yky
    59 cphy
    I think they are too few and too short to support hypothesis 1 above. For instance, if EVA:k corresponds to ‘l’, EVA:ockhy can hardly be considered a star name ending with -l (like in “Algol”) since k is in the middle of the word, not at the end (it possibly is the second sound, following Derek’s idea of the EVA:c-h ligature). Moreover, if suffixes were appended to the Arabic names, Voynichese names would be systematically longer than Arabic names, and this does not seem to be the case.
    So, if the phonetics hypotheses discussed on this site are correct, I think that the ending consonants are either dropped or changed into other sounds.

    What do you think is the most likely scenario?

    In my opinion, this point is also relevant to Derek’s observations about the possible grammatical function of the EVA:y suffix. Is that grammatical suffix already present in the Arabic words? Or is it a Voynichese addition / alteration?

    • MarcoP

      I am sorry: in the previous comment, I confused EVA:k and EVA:t.
      Ignoring the last and the second but last character, I count 9 star names ending in EVA:p/f/t (possibly corresponding to ‘-b’,’-f’,’-l’):
      2 odchecthy
      5 chocfhy
      11 cphocthy
      13 otys
      15 ocphy
      18 otol
      19 otor
      21 toeeodcthy
      59 cphy
      Actually, also an analysis of words ending with EVA:k and Arabic similar sounds would be interesting. EVA:k is also rare at the end of words (while many Arabic words, when “romanized”, end with -c, -k, -q).
      I hope my point is comprehensible despite the blunder above.

  30. Emery Fletcher

    With regard to star names (and their Arabic vs modern form), one well known star is the variable (eclipsing binary) Algol (beta Persei). It is a naked-eye variable, showing 1.5 magnitudes of variation on a two-day-plus basis (eclipse lasts 10 hours). The Algol name remains in use even among modern astronomers, and I believe such a notable star should appear somewhere in the Voynich star maps.

    It is most prominently visible around and after the autumnal equinox, and would pass nearly overhead for observers in south to central Europe. Despite its long-surviving connection with evil names (Demon Star, etc), I would think it would somehow be noted among the celestial pages in some manner.

    • PaulB

      I agree with you. Given that R and L are frequently interchanged in languages, what about star 10? Is it in the correct position?

  31. MarcoP

    Hello Stephen,
    I found this Persian globe that is not very ancient (19th century) but looks interesting.

    In particular, I think the Sun is represented on it (scarcely visible, on the lower left edge). If it really is the Sun, it would be interesting to know in which position it is, and if the Moon is represented on the opposite side of the globe (or is it the circle between Pisces and the hypothetical Sun?). A Sun and Moon on a globe would support the idea that the two 68r maps were copied from a similar object.
    It is also interesting that the globe only presents the images of a few constellations. Can you make anything of the names inscribed on it? I wonder if the labels are significantly different from the usual Arabic names. From the Flicker page, it is possible to download a large picture (it is just a little blurred, I guess because the globe was behind a glass).

    • Stephen Bax

      Thanks – the globe is upside down, and in any case the Persian names are tough to read! From my experience of similar globes, the names will probably be the Arabic ones. Can I ask: why do you say the globe has only a few constellations?

      • MarcoP

        Thank you Stephen, I had not even noticed that the globe was upside-down!

        I wrote that the globe presents only a few constellation images because the only images I can see are Pisces, the “sun” and the unidentified circle between those other two.

    • MarcoP

      Thanks to the help of user Eskandar on the Persian forum at unilang.org, I have been able to understand more of this interesting celestial globe.

      As customary, the globe is divided into twelve vertical areas of 30 degrees each. In the photograph, five such partitions are visible. Eskander has noticed that the second partition from the right is numbered and labeled “2 Twr” and the fifth “5 Asad”. “3 Jawza” is also readable in the central partition. So we can infer that the signs are labeled (from right to left): 1 Hamil (Aries) – 2 Twr (Taurus) – 3 Jawza (Gemini) – 4 Saratan (Cancer) – 5 Asad (Leo).
      The sun appears in the rightmost sign: Aries. This is consistent with Derek’s interpretation of 68r1: the Sun at the top marks the Spring equinox. Actually, I cannot exclude that the horizontal ring on which the Sun is engraved can turn around the globe (in which case the position of the Sun in Aries would be purely coincidental).

      I have no idea of why Pisces is represented far from ecliptic and in the sign of Gemini. The meaning of the double circle around Taurus and the single circle next to Leo is also unclear to me.

      • Stephen Bax

        Really interesting. Thanks to your clarification I can see the system now, and some words. One thing – 2 reads TH OW R rather than T OW R, the standard Arabic word for bull – for me it is interesting that it does not have the definite article AL in front of it, since that is the form as I read it in page 68r3 by the Pleiades.

        • MarcoP

          Hello Stephen, I see that the article AL is also missing in the Al-Buni diagram you linked here:

          Possibly the omission of the article was common in scientific diagrams?

          • Stephen Bax

            In my experience Arabic manuscripts usually use the article. In Arabic, the word without the article is odd, as it is singular, and would not be used for a proper name such as Taurus.

            So when I see it without the article, it makes me think it is a borrowing, as seems to be the case with the Persian globe and the Timbuktu manuscript.

            That is why it is interesting to me that the word which I read as ‘Taur’ on Voynich page 68r3 does not have the article, as it suggests a possible borrowing from Arabic, but not Arabic itself!

          • MarcoP

            Thank you Stephen! I had not realized that also the Al-Buni manuscript is an example of borrowing!

  32. Diane

    Marco, the assumption that the star-names will be the Arabic ones is just an assumption, not anything which becomes more true the more that people believe it.

    It seems to me that Prof. Bax, perhaps on advice from others, may be happily following up the work which I did (and of which I daresay he knows little) which shows a repeated and direct connection between the astronomical imagery and the needs of navigators. It was I who pointed out articles which showed that the early developers of what we term the ‘portolan chart’ included not only Genoese but north Africans and Basques, and then later researched the interesting question of when copper resinate first appears as a pigment in extant western art, since it had previously been used as a maritime coating. The research took me, not to England or Germany, but to the western side of southern France, if I recall, not far from Basque territory. Since I’ve also written and researched [in my other, more formal hat] about the distinctions between eastern and western sidereal navigation etc., the matter of mariner’s charts, pigments, stars and so forth is all of considerable interest. What I have come to doubt, however, is that any substantial part of the Voynich manuscript’s contents was first enunciated by a Latin-educated European.

    • Darren Worley

      I find the idea of a connection between Southern France and the VM quite intriguing. If the VM were a merchants codex (I favour a religious purpose) then I had a vague notion that it represented the work of overland merchants rather than maritime traders.

      Interestingly, an important group of Jewish merchants are believed to have originated in Southern France.


      These Jewish-Arab traders, the Radhanite, were involved in trade between the Christian and Islamic worlds during the early Middle Ages. Their trade network covered much of Europe, North Africa, the Middle East, Central Asia and parts of India and China.

      Surprisingly, the start of this Jewish trader route is believed to be in Southern France. The name Radhanite coming from the Latin “Rhodanus” refering to the river Rhone. These merchants are said to have spoken Arabic, Persian, Roman, the Frank, Spanish, and Slav languages.

      Here is an excerpt from a 9th-century Muslim source.

      Quote : The merchants that start from Spain or France go to Sus al-Aksa (in Morocco) and then to Tangier, whence they walk to Kairouan and the capital of Egypt. Thence they go to ar-Ramla, visit Damascus, al-Kufa, Baghdad, and al-Basra, cross Ahvaz, Fars, Kerman, Sind, Hind, and arrive in China.

      They seem like a good candidate group in the transmission of Judaeo-Arabic/Arabic texts into Europe. Presumably in addition, to trading in “eunuchs, female slaves, boys, brocade, castor, marten and other furs, and swords” they might also have traded in manuscripts and ideas.

  33. MarcoP

    Hello Stephen,
    I have been thinking about the fact that 50% of the star names (33 out of 64) end with EVA ‘y’. This does not seem to match the distribution of Arabic star endings: no single letter has such a high frequency. So, while the high frequency of star names starting with EVA ‘o’ can be explained by a derivation from the Arabic article “al-“, I cannot think of an explanation for all these occurrences of the suffix “-y”.

    I guess you will address this in the forthcoming statistical analysis you mention at the end of this post, but could you anticipate your ideas about this specific subject?
    Are we facing some kind of “systematic corruption” of word endings that happened when the Arabic star names were assimilated into Voynichese?

    • Stephen Bax

      I’m sorry my statistical discussion about stars has been delayed. As for your interesting observation – to the point as always! – I some time ago came to the conclusion that the EVA ‘y’ symbol at the end of words is so systematic that it is probably some sort of grammatical marker (e.g. an indefinite or definite article) or else a marker of gender, or something else which is not part of the root word.

      This seemed to me to be the case with the famous ‘Taur’ word on page f68r3, and it seems to me to be systematic through the manuscript. Your analysis of its appearance in the star names would seem perhaps to confirm it.

      So for that reason you might notice that when I try to transcribe the possible star names I usually miss it out altogether. But I could be wrong!

      • MarcoP

        Thank you for the clarification, Stephen! Since the current hypothesis is that these star names are proper nouns, the -y suffix appears particularly mysterious in this context: very interesting! Thank you also for your suggestion to tentatively ignore the suffix when looking for correspondent Arabic star names: I will try to reconsider a few star labels according to this approach. Maybe, if a sufficient number of stars can be reliably identified, it will be possible to understand more also of this particular issue.

        • Darren Worley

          I came across the following reference in a book on the Dead Sea Scrolls, and thought that it could be relevant to understanding the origin of the “y” suffix that is frequently found on VM star names.

          According to the phonetic mapping, EVA:y represents the sound “n”

          Quote: “..a number of names found in Merkabah Kabbalistic literature adopt the suffix -on”

          The footnote suggest that this is taken from the book Kabbalah, by Gershom Scholem [pub; Penguin, 1978]

          I also wonder if the prefix that is often found at the beginning of VM star names, might represent “ha”, the Hebrew definitive article, rather than “Al”, its equivalent in Arabic?

          • Derek Vogt

            I just ran across these posts on the EVA-y endings again, not long after noticing a pattern in its use in star/constellation names that I’ve tried to match one-by-one on 68r1. That final letter tends to show up on names that refer to something which the named character or object does: driving a wagon/chariot, kneeling, barking, landing after flight, carrying water. That makes it look as if it were equivalent to our “-er” and “-or” suffixes.

  34. Paul Hicks

    Is the star above pisces f70v2 Al Kaitain = Alpha Piscium?

  35. Sorry if this has already been noted elsewhere, but I just came across an image of a “Late Babylonian (Neo-Assyrian) planisphere” and its similarity to the third star image struck me as remarkable.


Leave Your Comment

Your email will not be published or shared. Required fields are marked *

You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <s> <strike> <strong>