Mamillius, an Aldine book type

George Horton's picture

This Aldine lowercase has been partly repaired after disastrous alterations, but I no longer think it has sufficient intrinsic character to be worth going on with.

Current Mamillius.pdf9.69 KB
Chris G's picture


this looks good, the forms are very lively. I like the 'g' a lot

I'd alter the top of the 'e' as it looks as if its leaning to the right a little, and adjust the baseline/x-height overshoot of the rounded forms such as 'o' 'e' etc. At the moment they appear to float above the baseline. Similarly the alignment of certain letters to the x-height is a bit variable, see for example the 'ng' and 'wn' combinations. The terminal of the 'j' could benefit from an approach closer, but not identical, to what you've done with the 'y' as it seems a little unsure of itself at the moment

hope this is of some use

I think the finished article will make for very comfortable reading.

George Horton's picture

Dear Chris,
Thanks very much for your help - your comments were perfectly accurate. A revised PDF is up.

point918's picture


The old-face irregularities in this font give it a pleasant warmth and – I agree with Chris G – make for a comfortable read. I would counsel against over-polishing it, as digital tools tend to encourage. The one thing that caught my eye (in a negative sense) in the test setting was the y. I wonder if the counter isn't just a tad small and contributing to the sensation that the long diagonal is too long ... in detail the terminal also seems tentative; unresolved in the context of the decisive and crisply cut serifs elsewhere in the face

George Horton's picture

Dear point918,
I've warmed it up further by reducing the concavity of serifs - discarding what is only a gimic without Baroque letterforms. But this tends towards, not away from, digital regularity. In fact I don't want irregularities in the sense of a pseudo-random scattering of yokelish accidents. Rather, I want to avoid constructivist letterforms.

You're right about y, I've enlarged the counter: thanks.


Miss Tiffany's picture

The problem with doing any sort of a revival is how to introduce your own ideas into it. Without introducing your own ideas into the design you risk being to obedient to the original and then you are stuck with the problem of lack of originality.

I wouldn't remove the concave shapes, rather I'd figure out a way to introduce that elsewhere.

George Horton's picture

Miss Tiffany,
This isn't a revival of any design but rather a type developed independently, within the framework that Griffo invented: the motivation behind the letterform is roman-calligraphic, but this motivation is expressed through typographic marks, including typographically shaped serifs. Rather than ideas I have brought my taste to the design. Originality hasn't been a aim, though mamillius does have a character in text different to those of existing types. And I think those concavities were pretty gross.

Randy's picture

Hi George.

This is really well done. On a structural level, the proportions and color look good. Here are a few areas to take a look at:

h,m,n: the right legs seem a tad to bow-legged for my taste.
z: bottom is slightly long
y: agree with Point's above comments
decenders: slightly too long I think. (will help the y too)

The next thing I think you should work on is the quality of the outlines. On the whole they look wobbly. You miss this at text sizes where things look fine, but look to the terminals of the s to see what I mean. Here are a few areas to look at:

s: The spine looks fine. Nice and smooth curves. It gets ugly from there. The terminals (especially the top one) are very uncertain. Pick a line and go with it :-) Maybe look at your f/j terminals or z for inspiration. The curves in the counters are bad... too square. (Perhaps you adjusted the overshoot by dragging down the top points without adjusting the beziers?) Rework them so they flow with the same grace as the spine.

t: upper left is wobbly

There are many places where I expect a single curve, but find an s shape.

You have three kinds of serf endings. Why? Some of the head serifs have totally rounded tips (p), some have corners and bowed sides, while others have corners and flat sides? You don't pick this up at text sizes but these are the kinds of things that make a quality face IMHO.

I wouldn't mind seeing more pen in the punctuation ie not perfect circles.

All that said, this is very well done.


George Horton's picture

Hi Randy,
Thanks for your sharp comments. I've updated the pdf in partial agreement with them.

s and z have been altered. the top terminal on s was a late-night accident, but the principle of uncharacteristically sharp terminals on s is I think defensible - though I've now regularised them. You were right about z. I'm sticking with long descenders: it's a slippery slope down to ITC Garamond once you start shortening descenders. t I have grudgingly unwobbled a little.

The serifs are now a bit more regular. I think they were more continuous in their variety than perhaps you noticed - all head-serifs, for example, were rounded, though some more than others - but, as you say, they now make a better impression at display sizes. Bow-legs are staying. Punctuation and dots on i/j I'm still not decided on. Any ideas as to how it's best done?


Randy's picture

I like what you've done. I still wonder on the little flatop on the head serifs. It seems like a sub visible design element until the type gets to a size where this*feature* becomes odd. It just seems out of character. I don't see it being a result of ductus or of function or of style.

Decenders: You've set with generous leading, because you must to prevent clashing. If this truely is for extended reading, consider heading down the slippery slope about 15%. It will still be elegant I promise.

r: Did you modify this since last time? Hard to tell since I deleted the pdf I downloaded before. Anyhow, I like how you've dropped the connection down (compared to an n). And I like how you've tried to introduce more white space by raking back the head serif. I think you've overdone it a tad though. Consider bringing it back touch.

Punctuation: Show some ductus, ie more diamond, less circle. It should reflect the angle of the *pen* used to create the leg of the n. It seems like slightly rounding the corners (or bowing the sides) will fit nicely with your serif treatments. You might start with the comma. Instead of a dot with a hook, try an arc as created by your *pen* And keep it nice and long. One critique I've received and seen repeatedly at critiques from Matthew Carter and Akira Kobayashi is "make the comma longer for text."

Looking forward to caps and italic. This is a very elegant face. Well done so far.


George Horton's picture

Thanks again Randy; as you can see I have followed your advice, with good results. I've learnt a lot over the last few days, and for the first time I feel I know what I'm doing. It's now very satisfying work.

Randy's picture

No doubt. It's looking good. Further thougts:

y: Still don't like the tail. I think the tail is maybe a little too thick too close to the crotch. More of a flair might help, or some kind of foot, or terminal like your j.

punctuation: I like the ductus! I wonder why the dot on the i and j don't also reflect this change? The following is nit picky stuff: I like the diamond period (though it seems a little squished vertically -- like it wasnt' drawn with a pen at 45degrees in one stroke, upperleft to lower right. The comma is too angular imho. Try a more vertical hook (don't worry about keeping diamond shape at the top). This will have two benefits. Better spacing with letters like g. Also stylistically it works with your bowed nhm. Not hard angles in this whole font!

It looks like you know what you're doing!


George Horton's picture

Hi Randy, good of you to check again. I've done pretty much as you suggested, though you might prefer more drastic work on y, and the comma might be less hooked.

Randy's picture

1. I found the old pdf in my cache and prefer the old r (to your re-revised r) -- do like the long foot.
2. Perhaps some more heft in the the Terminal of the f
3. j: the bend is a little thick. Maybe take some off the outside (see the f)



George Horton's picture

Thanks Randy, I'll get going on the caps soon.

seml's picture


You may find this odd, but I believe this typeface looks a blend between the Aldine, as you refer, and a Neoclassical type, should I say, a baskerville in some details? It does remind me of Bembo, but with a touch of Baskerville.
It's well conceived overall, but I would like to see the uppercase and ligatures (are you developing them?). And I also believe an Italic would be interesting, but you must have thought of that ;).

Curious, but you don't notice how long the serifs are. Only when you zoom in, which makes it a good baseline.
As for particular shapes, I would need some more attention.

Keep up the good work.

George Horton's picture

Hi Seml,
The bias of curves in Mamillius is blander and less mannered than in Bembo - perhaps that is what reminds you of Baskerville. Uppercase and ligs are being made, yup, and either a Joanna-style italic or a digitisation of Griffo's superb italic for Soncino (link).


jafo's picture

Hi. Another voice of approval for the design and warmth here. It feels organic, like twigs growing unhurredly in the summer sun. I like how you've reinterpreted the classic Aldine by harmonizing it with notes from the endearingly wobbly Goudy; I particularly like its subtle, angular Jenson (Centaur, really) influence. Overall, it's a rather successful design, modulo a few letter sequences: the round, slightly bowed h/m/n combined with the angular e are stylistically jarring, for example -- Bembo vs. Centaur. But the overall effect is excellent, and I'd really like to see the UC and italic.

George Horton's picture

Thanks Jafo; I've just cleaned up some errors. Very perceptive of you to notice Rogers rather than Jenson, and there's lots of Goudy here, in the s and a especially - I admire his work greatly.

George Horton's picture

Postscript: there's not much Goudy left in a or s, and the curves are less blandly neoclassical, but I think the textblock has a clarity now that it generally hasn't had during the very long development of this lowercase.

.'s picture

Dunno if I'm repeating anyone else's comments, so many apologies if I am.
The descender of the "y" gets pretty heavy. Maybe the right side of it could be taken in a little bit?
The bottom of the "g" looks a little flat; maybe tweak the curves leading to the point to give it a bit more curve.
The "j", "b", and "l" look a little heavy in context. Don't be afraid to make some of your verticals lighter than others; there are some "tricks of the light" which should be compensated for, even at the rsk of consistency of construction. (I have discovered this on even very constructed sans serif fonts.)
The "v", "w", and "x" look a little light in context.
The top of the "a" seems to be overreaching, but it works very well. I like it a lot.
The cap "I" looks light in context. Caps benefit from some additional thickness.
Keep rocking.

ebensorkin's picture

As a preface let me say that I too am impressed. I am naive about text faces still, and will be for some years yet I suspect. So take my crit/questions with some ( or even a lot of ) salt!

So far all the comments I have read have dealt with letterforms rather than interletter spacing. Is it the case that eveybody is happy with that aspect of the font or is that being saved for later? When I look at the word 'child' the spacing does seem to be an issue. The l seems to be clinging to the i to escape the d. The w seems a bit over strong. It takes over. The c seems to be creeping to the lower left. The spacing of 'ya' in 'yard' seems to big. Overall it has the effect for me of reminding me of metal type that was set without kerning. The i in metal Bembo seems to creep to the a on it's right and away from an n on it's left. I wonder if philisophically you may be looking to relate closely to that history & approach.

Seperately, the tail of the 'y' seems to draw a little too much attention because of the weight. The 'a' does too but for the opposite reason - it's terminal is so light.

The irregularity of the face does make it lively and pleasant ( other folks said organic, wobbly, pleasant warmth etc ).

I also think that the ascenders & descenders could be less tall & deep with no ill effect.

It would be nice if you posted the old PDFs so I could see the progress as you went along.


George Horton's picture

Chester and Eben, thanks for your comments, I've acted on them but I think this might still need more thought. See what you think now - it finally looks OK at 15 point, though its colouring is rigidly 11-12 point.

Kerning is not yet done - and yes, I do love metal Bembo's slightly odd but very intelligent spacing.

I haven't kept old PDFs, or indeed dated backups. This has resulted in my having to go with whatever tweaks I've made, typically with horrendous results, but it does mean that I've been able to learn a lot about the effects of small changes on the textblock.


ebensorkin's picture

> I haven’t kept old PDFs, or indeed dated backups.

How extraodinarily naughty! You could change your habits now of course...

As before take all of this with the understanding that there are lots of folks here who know far more than I but with that said:

I was thinking about a face like this and it's connection to metal type and thinking that it would be a good idea now to decide if you are going to make different designs for different sizes, something I think would add to the value of the project, and if so pick a size to work into first. Then you can make it behave for that size and not try to straddle to many purposes at once.

I also think that it's time for you to start thinking about interletter spacing and shapes at once. Even though you have been doing a great job in general there is just as much to accomplish in the spaces where letters are not as in where they are. And I don't just mean the bowls & counters. And I don't just mean kerning, I mean that you should design the spaces that occur in between the letters. It's a different way of looking at the whole thing. Some folks would say that letter shapes should be done with those spaces in mind. I would be in that crowd.

The used to be a nice teaching flash file on Typophile from Jonathan Hoefler that dealt with this. But the link to it is broken.

If you want I'll point you at threads that deal with this idea & other resources.

Also, Have you been here?

It's pretty cool.

George Horton's picture

Eben, it's OK, this is Mamillius 12 point, and will never appear at larger sizes if I have anything to do with it. I have indeed been thinking about intraverbal spacing a lot - that's why the head-serifs are how they are - what irks you about the current setup? l was loosely spaced to the right, but that's been changed.

All links are very welcome. I'm not sure I trust Hoefler, who is, let's face it, better at display than text faces (Requiem Text is rather flimsy, and the letters in words are polite to one another rather than having fun with one another). I do know about, but it's pretty basic and can (& did for me) give the wrong idea about eg relative stem weights - it's no good deciding how thick a thin is before drawing all the letters.


hrant's picture

> will never appear at larger sizes if I have anything to do with it.

Only if you'll be the only person to have anything to do with it...
And maybe not even then, since people change their minds. What I'm saying is, never say "never"! :-) That said, it is indeed sometimes useful to focus a design on a small size range.

> the letters in words are polite to one another rather than having fun with one another

Good point.

> it’s no good deciding how thick a thin is before drawing all the letters.

But how can you make any letters without deciding that?
The trick is to see any decision as provisional.


George Horton's picture

Hi Hrant, I doubt whether this would ever be used by anyone but me and a few friends, I'm afraid. I hope to do the MA in type design at Reading next year, and to immerse myself in Griffo sufficiently deeply to produce a really good Aldine type then, but for now this is both a chance to see how far I can go without training and a type for my own first book on Shakespeare.

Until yesterday, this design looked terribly dull when (as it were) blown up beyond 13 point. Checking how a face looks a little larger than its intended size can show clearly some kinds of structural failure only vaguely and nigglingly apparent at true size.

And what I should have said was "in particular, the weight of thins is not a design decision which should be imposed on letterforms in advance of their being drawn, except in the roughest terms". Which is rather different to, though much less silly than, what I in fact wrote.


hrant's picture

> and a few friends

Well, there you go - out of your control! :-)

> Checking how a face looks a little larger than its intended
> size can show clearly some kinds of structural failure

On the other hand, I personally feel that a "real" text
face has to have a certain ugliness when viewed large.

> the weight of thins is not a design decision
> which should be imposed on letterforms in advance

If you mean what I think by "imposed", I would agree. But the more experience one has, the closer one can come to establishing the desired parameters from the get-go; sometimes close enough not to need revision. It is reputed for example that towards the end of his career E Prince used to cut his finished punches with zero corrections, and IIRC without even doing smoke proofs!


George Horton's picture

It's reassuring to hear that a real text face has to be ugly when enlarged. Do you have an opinion on whether Mamillius' structure is exaggeratedly or sufficiently "text"?

ebensorkin's picture

About the size issue. If you have picked a size you are designing into then thats great. My sense was that you were thinking about evaluating the type at several sizes and so I started to think that you were attempting to straddle 9, 12 & 15 or something which could only lead to misery. I see now that is not your intent. Chris Lozos recently said that rather than working on multiple weights (extra thin, thin, bold, extra bold or whatever) of a font it might be better to start by making the same font but cut for several intended use sizes. I think he is right about that. Even if you are designing for just one specific size.

About relative stem weights. No source is completely oracular. Casting about for a source that dealt with the idea of interletter spacing that site was the best I could find quickly.

Certainly there is a problem of overpoliteness in lots of digital faces. Adobe faces seem to exemplify that. And certainly you are well on your way to avoiding that problem.

I am not inked by your type by the way. It's pleasing to me. It just lacks a certain virtue I would love to see it have. It's a virtue that is pretty rare and one that I am still developing my eye for myself.

Because I am after a little more harmony in interletter spacing, some of what I am after could be construed as suing for that same well mannered but overly polite quality. But I don't think you have to sacrifice color & warmth ( aesthetic virtues) or readability ( a use virtue) to achive better harmony.

About this spacing issue you said: that’s why the head-serifs are how they are.

It isn't the serifs that concern me but the actual amount & shape of the white space between glyphs. Not the black but the white. Look at the word 'was' The w is drifting away. You could say but the serif... but that isn't the issue. The issue is the design of the w ( and a), and the design of their relationship - the design of the interletter space. Theoretically the design of a given a or w could be perfect. As long as the other letters harmonize with it. But it is easier I think to design the harmony in the first place.

I don't want to say make your 'a' like this or 'w' like that because that just deals with one or two specific places where something is happening. If I convince you at all ( which i might not ) I hope it will be to look at the type and judge it in an additional way.

Hrant has influened my thinking about this a great deal when he wrote about Legato. (Scroll way down)

He & others have also talked about this idea and related ideas using terms like 'bouma' and 'notan'. You could seach for these terms on typophile & see what you think.

Here is what the designer said about Legato

>On the other hand, I personally feel that a “real” text face has to have a certain ugliness when viewed large.

I agree. Features that are going to be preceived well at all at 12 point will become grotesque when large. And a display version will look weak-willed at a small size.

hrant's picture

> It’s reassuring to ...

Woah, don't be too reassured! :-) Many accomplished type designers do not agree with that idea. These tend to be the same people who don't place a qualitative distinction between display and text fonts, believing that it's just a matter of changing stroke contrast, etc. to get one to be the other. But to me these are the designers who produce non-optimal (if still entirely serviceable) text faces.

> Do you have an opinion on whether Mamillius’ structure
> is exaggeratedly or sufficiently “text”?

No, I'm not good enough for that yet. Although the statement I made I have full faith in, I'm still not at a point where I can see exactly what it is and make it happen myself; which is why Patria is still too close to the Display end of things - although some of the people described above still complain that it doesn't have enough formal harmony. What they don't grasp is that that's a feature, not a bug.


George Horton's picture

Eben and Hrant, this is interesting. I agree that Mamillius is missing something, no doubt several somethings, and I think one is white "rythm". But what genuinely accurate phrase should replace "rythm" there? An adequate name for it would also be a description of it. Smeijers' 'Counterpunch' has the beginnings of a description of white harmony through the reuse of counterpunches. But presumably there's a whole world of different kinds of white interractions, just as there is of black, and a big part of the problem for the novice is that the shapes involved are so different (that most are composite is a smaller and more obvious difficulty).

hrant's picture

> what genuinely accurate phrase should replace “rythm” there?

I don't use "rhythm", because I think it alludes to a motion that
is not there. I prefer "pattern". And to me the ideal objective is
to achieve a certain pattern in the notan (unity and relationship
of Black and White), the desired pattern depending on the purpose
of the design.


George Horton's picture

OK, I've started working on the notan, and already there's a lovely sprinkling of fairy-dust on the textblock.

ebensorkin's picture

George, I am glad it's helping!


Hrant, explain to me why the the motion is not there.

Patten in motion is rhythm - no?

The eye moves past the pattern we already agree is there. So...

hrant's picture

> Patten in motion is rhythm - no?

Exactly, but where the motion is continuous. Saccades preclude that; mechanically, reading is an aggregation of discrete fixation-events. Now, some people will tell you that rhythm does not depend on motion; but to me, whatever their motivation*, they're causing trouble by using the term.

* Quite often a stubborn/romantic apologism.


ebensorkin's picture

Hmmm I am not convinced about this yet. Rhythm which is regular in nature as in 'musical' would probably need to be constant or regular which is what I think you mean when you say 'continuous'. Continuous might just mean continuing with varied pace. But rhythm can also be irregular. Obviously this is bad in a heartbeat, and probably alot of music but other natural events have more complicated structures - like reading. But that irregularity & complexity don't preclude rhythm in my mind.

> Now, some people will tell you that rhythm does not depend on motion.

That sounds pretty specious to me!

If somebody made a claim for visual rythym and wanted to allow for a wide variety of motions or complex motions a opposed to a single motion I would be willing to think about that.

> reading is an aggregation of discrete fixation-events

Light is considered both a particle & a wave. Reading is not light (obviously) but I am not sure that fixation events which occur between saccades preclude rhythm. Also, saccades are movement. Even if there is no motion/rhythm during a fixation event, something I am not sure is true yet*, there is the saccade itself.

What makes me skeptical of the idea of rhythm isn't the objections you have brought up so far. It's the issue of 'frame'. In musical terms the 'frame' is very narrow. A moment of conciousness. In reading I wonder if the area of focus isn't too broad for the term to apply well.

So am am ultimately ambivalent about the term.

> Quite often a stubborn/romantic apologism

Romance is stubborn. It makes me laugh happily when you write stuff like this. ;-)

Look, we are hijacking George's thread.

George, I think your term 'interractions' is decent and covers the needed idea. But Hrant's term 'Notan' seems the richest and most satisfying to me for many many reasons not least of which because it is a term which is all about looking, noticing, and making your own judgement - a responsibility that can't ever be shirked in favor of typographic dogma of any stripe.

*Hrant, If I am going to keep argueing with you about this kind of stuff I had better start reading some primary scientific literature!

George Horton's picture

Hrant and Eben, I am enlightened, see what you think of this now.

George Horton's picture

Sorry about the following empties. Can they be deleted?

ebensorkin's picture

The font doesn't seem to be fundamentally altered, but how could it be this quickly? That said I think it's better.

I must say that it's a good thing I have kept my old print outs or I could not even say that much. Your habit of overwriting the old PDF makes it impossible for other folks to follow along & benefit for your process afterwards. Which is of course VERY naughty as I said before!

What I notice is that the words seem to gel better now. They seem less restless. The eccentric qulatity of the c and e ( the wayward lower left) is not as noticable now.

The 'w' still seems like less of a team player than I would ideally like. It seems too wide, too tall, and to plunge below the base line too far.

The 'a' is in contrast, a bit to much of a shrinking violet. I keep wondering if the top needs to take a bit of the pluck of your 'r' or if the bowl might be bigger, or maybe both.

What a is your a modeled after?

What do you think? What have you noticed?

George Horton's picture

Hi Eben, yup, I've been concentrating on the gelling of words, through the resizing of counters and also by rotating white space counter-clockwise relative to the axis of black. I've found that Griffo and, to an almost exagerrated extent, the draughtsman of Monotype Bembo observed the same dual-axis principle. It gives sparkle despite low contrast, which is now the most notable thing about this design, and it helps, as in FF Legato, to unite the word-image.

The unterminated 'a' is found in Jenson's and Gill's types, and also in Lutetia.

Sorry about not posting old PDFs, I've put up one previous version along with the current update.

ebensorkin's picture

How interesting! I'll have to go & look at those examples to get some better context.

I am looking foreward to seeing the next PDF.

George Horton's picture

Hi Eben, more thorough changes have been made, and a new PDF is up.

ebensorkin's picture

I have printed out the two samples. It will be a day or so before I can get into it. I just saw some type in a fancy edition of Moby Dick that had more than a passing resembalance to your font. Similar shapes & spacing. It was pretty weird. I'll look into what it was & let you know. I should have jotted it down at the time but ...

Also I need to look at goudy centaur & bembo again too so I have a better idea of where you are coming from. These are the big three sources you are referencing - no?

George Horton's picture

Actually I'd leave it for a while - I'm still not very happy with this.

ebensorkin's picture

Hmmm- okay. In the meantime though, it is goudy centaur & bembo I should look at by way of comparison. Yes?

Also, what specifically are you not happy about?

George Horton's picture

Sorry to fiddle around Eben; I was overwhelmed with disgust at Mamillius' bland ugliness, but I think I was just bored with these now rather overfamiliar letterforms - a fresh eye would be very welcome. The copy you will have printed out is thus the correct one. Thanks very much for taking the trouble to look at it,

George Horton's picture

- And the relevant fonts are Bembo Book and Poliphilus.

Syndicate content Syndicate content