Clumping of Non-Granulating Watercolour Paints, Part 2 – It’s not me, it’s you, M. Graham

Several weeks ago, I wrote what turned out to become my most popular blog post yet.  In it, I explored the effect of water hardness and dissolved mineral content in water on watercolour paints on the palette and on paper.

In particular, I was trying to find an explanation for some weird curdling/clumping behaviour I was observing in quinacridone paints from M. Graham.  When I experimented by switching my brushwater from Kitchener-Waterloo tap water (some of the hardest in North America, containing extremely high levels of dissolved calcium and magnesium) to store-bought distilled water, I observed a marked reduction, but not a complete elimination of curdling effects

Original Experiment:  Tap water vs. Distilled, used pans/washed in tap water

At the time of my original post, I attributed the remaining curdling/clumping in the distilled water sheet to mineral buildup in my pans, brushes and palette from previously using tap water.  I theorized that if I switched to distilled water for my painting, over time, this would improve.  I could also reduce the effect of mineral residue on my palettes by wiping them off after washing, and rinsing my brushes in distilled water.  At the time, I was cautiously optimistic that I would be able to all but eliminate this problem with changes to my painting routine.

Over the past few weeks, I have been dilligent about using only distilled water. I no longer think this is a realistic solution, and I now have serious  doubts that this is the only reason for my M. Graham paints curdling.  A few things have changed my mind:

  1.  Despite my dilligence in using only distilled water, all the affected M. Graham paints continue to curdle.  In the case of the quinacridone rose, this effect has reduced enough that it is now no longer as bothersome.  However, the neutral tint still sometimes separates, and the Quinacridone Rust(PO48) still clumps dramatically.  In this case, I have gone as far as to fill a fresh pan, tried using paint straight from the tube, even washed my palettes with distilled water, to no avail.  No matter what I do, if I water this paint down, or try to mix it with another colour, it still clumps. If this is indeed simply a reaction with some minerals, it is reacting with extremely low mineral content, lower than most tap, filtered or bottled water most artists use.
  2. While I am willing, grudgingly, to buy distilled water and haul it to the studio for studio use, I also enjoy carrying around palettes for urban sketching, field studies, etc.  In these cases I will use whatever water is handy – filled from a drinking fountain, lake water, bottled mineral water, etc.  I now find myself concerned not only about whether my paints will unexpectedly clump with the water of the day, but whether I will be messing up my pans by leaving mineral residue in them.  This is ridiculous and unacceptable
  3. In a bizarre mix of frustration and optimism, a couple weeks ago I purchased both a replacement PO48 (Quinacridone Burnt Orange, by Da Vinci) and another supposedly non-granulating M. Graham paint (PG36- Phthalo Green YS).  The Da Vinci paint performs beautifully, a very similar colour to the M. Graham, but not a hint of clumping no matter what water I throw at it – whatever the issue is, it’s clearly not merely pigment related.  The new phthalo from M. Graham, on the other hand, does some bizarre stuff.  It clumps dramatically on the palette, although how much varies day to day, but then the clumping relaxes and all but dissappears as the paint dries on the paper.
  4. Just as I was beginning to really think I was going insane, consistently seeing paint behaviour that nobody else has reported, an amazing Youtuber I follow,  Sadiesavestheday, commented on the same thing in one of her videos (skip to 14:30)

I’m running out of ideas of what could be causing this.  It has occured to me that it could be the honey crystallizing inside the paint due to living in a climate with dramatic temperature swings, but that would also fall under the unnacceptable paint behaviour category.

So where does this leave me?  I’m actually really upset, because, as I’ve raved before, I love the rewettability and saturation of M. Graham paints.  I love their pigment choices and focus on single pigments. Their PY150 Nickel Azo Yellow is hands down my favourite (read: only) yellow. and I love a bunch of their other pigments too.  However,  at this point 4 of the 10 or so M. Graham paints I own display some level of frustrating clumping.  While it is possible to mitigate and work around this in most cases, none of my paints from any other brand have this issue, so I am reaching the conclusion that it will simply be easier to just buy from other brands in the future.

So I guess this is my breakup letter to M. Graham, at least for now.  M. Graham, it’s not me, it’s you.  I just can’t handle your high maintenance needs and mood swings anymore.  I’ll be moving on to more reliable, easygoing paints.

I will reach out to the manufacturer at some point in the near future.   I hear they are very receptive to feedback. I hope that they will be able to track down the root of the issue and fix it, so I can once again enjoy these pretty colours.


The search for transparent yellow

I’ve realised that not only am I generally a fan of transparent,non-granulating pigments in watercolour, I am particularly fussy about this preference when it comes to yellows.  I can enjoy a fair bit of granulation and opacity in mossy greens and browns (great for adding visual weight to heavy foliage, bark and rocks).  The opacity of Cobalt Teal is what gives it that absolutely yummy irresistible bright colour in the pan.  Softly granulating, semi-opaque manganese and cobalt violets are great for laying in shadows, and even one of my favourite primaries, Purple Magenta (PR122) is far from transparent.  But even the smallest hint of milkyness in my yellows makes me gag.  In my ideal world- all yellows would be like clear, pure liquid sunshine.

This is a stroke of terrible luck on my part.  Whereas transparent pigments abound in some sections of the colour wheel – there’s a transparent, non-granulating phthalo pigment for every shade of blue-green you can imagine, transparent yellows are few and far between – the most common yellows being opaque cadmiums or semi-transparent azo and hansa pigments.

As a result, for the past year, I have basically used one yellow pigment, PY150 (Nickel Azo Yellow or Transparent Yellow) for all my yellow needs. I do also use small amounts of Green Gold (PY129) and Quinacridone Gold (PO49), which are arguably also part of the yellow family, but PY150 is the only “true”  yellow I have used in a painting in the past year.

Nickel Azo
My single yellow – PY150 Nickel Azo Yellow by M. Graham

As an unapologetic pigment magpie – at last count I had 15 blues in my half-pan palette, most of which I find a use for regularly, my yellow section has been relatively anemic.  I do keep some Hansa Yellows PY3 and PY97 I tried and never took to, along with my trusty PY150.  On somebody’s recommendation, at one point I also bought M. Graham’s PY110 Indian Yellow, which is a lovely colour, but also rarely used (and really more of a yellowish-orange, IMO).

Apart from an impulse to try ALL the things, having only one yellow I regularly use hasn’t been much of a problem.  As a botanical artist – I primarily use yellow to mix greens for leaves. PY150 is very close to a primary yellow in tints, and I have so many blues and greens I can easily mix any green I could imagine with only the one real yellow.

However, the trouble with PY150 as a sole yellow comes when rendering pure yellow subjects such as flowers.  Transparent pigments are darker in masstone – in the case of PY150, this masstone shifts to a greenish-brownish sludgy colour I like to refer to as “birdshit brown”.  In rendering leafy subjects, this is never a problem, as this colour works with the shadows.  However, pure yellow flowers often want similarly “clean”  shadows in a warmer orange colour or a cleaner greenish one.  So a few months ago, I started once again feeling the itch to try exploring alternative yellow pigments.

Reading up on pigments, I became very curious about PY153 Nickel Dioxine Yellow, often sold as “Indian Yellow”  or “New Gamboge” up until 2012 or so.  It is described as transparent, non-granulating marigold yellow in masstone fading out to a warm primary yellow in tints.

Manufacturing of PY153 ceased a couple years before I started painting in watercolour. Most lines have since discontinued  the once popular PY153, but a few brands still use it and of course there’s some older tubes still floating around.  I decided to try to find as much PY153 while I could, and stock up if I liked it.

From Jackson’s Art Supplies in the UK – I ordered two PY153 paints.  Sennelier Yellow Light, still made with single pigment PY153, is an oddball formulation that is much lighter/greener than the Indian Yellow/New Gamboge usually associated with the pigment.  I also ordered Jackson’s own brand Indian Yellow, which claims to be made with PY153.  I actually question whether this is true – Jackson’s brand is also made by Sennelier, who a few years ago reformulated their own Indian Yellow to be a mix of PY153 with PY154.  However I will trust the label, as it is possible that it is older stock, or that Sennelier is still making paint they resell to other labels with the old formula.

I also wanted to try some of the popular, discontinued Winsor Newton PY153.  Luckily for me, I live in a smallish town – large enough to have art stores, small enough that their stock turnover rate is extremely slow, and was also able to get my hands on some Winsor Newton Indian Yellow made with single-pigment PY153.  Winsor Newton also apparently sold PY153 as New Gamboge, although I could not find any left over in the local shops.

Comparison of “Transparent” Yellows  – 3 Formulations of PY153,  3 Indian Yellows, PY150

In this way, I’ve managed to collect 3 different formulations of PY153.  I also, rather confusingly, now have 3 paints named “Indian Yellow”.  In the test swatch above, I’ve painted out Sennelier Yellow Lt (PY153),  MG Nickel Azo Yellow (PY150), Winsor Newton’s old formula Indian Yellow (PY153), Jackson’s Indian Yellow (PY153) and MG Indian Yellow (PY110)

On the bottom line I glazed all the yellows over Winsor Blue GS – clearly I didn’t let the latter dry long enough as it bled into the yellow paints.  It is a hot, humid day here.

My results were somewhat mixed, and I feel somewhat foolish for spending a whole bunch of money on yellows when really I’m happy painting with PY150 99% of the time

PY153 Sennelier Yellow Light (Sennelier)  is a beautiful pure yellow colour, just to the lemon side of primary.  It would make a good addition to my palette, but it really isn’t transparent.  You can see the yellow over the top black line.  Also, annoyingly, it is less transparent when wet, which really adds to my gaggy reaction painting with yellow milk.  That said, for those who don’t have such an illogical response to non-transparent yellows, it is a pretty colour.

PY150 Nickel Azo Yellow (M. Graham) remains my favourite yellow.  The birdshit brown really doesn’t look so bad unless it’s REALLY concentrated, and it makes a beautiful variety of greens and oranges in mixes.  I would probably still choose to paint most things with this.

PY153 Indian Yellow (Winsor Newton) This is my favourite of the PY153 paints (and of the paints named Indian Yellow).  It is the most transparent and the most chromatic of these groups, although still less than PY150 – ranging from a golden orange to a beautiful marigold yellow, down to a buttery colour in tints.  It is really too bad this is discontinued.  I’ve managed to get my hands on 2 5mL tubes.  I can see reaching for this paint to paint warm yellow flowers such as sunflowers or black-eyed susans, where it would capture both the body colour and midtone shadows perfectly.

PY153 Indian Yellow (Jackson’s) This is similar to the WN paint but slightly less transparent and slightly less chromatic.  I don’t know if this is down to the paint manufacturing, or whether the Jackson’s tube is incorrectly also contains the less transparent PY154.  On the brighter side, it is about a quarter the price, not discontinued, and comes in massive 21 mL tubes – a couple of those should last a lifetime.  If you are less fussy about transparency, it’s a nice colour.

PY110 Indian Yellow (M. Graham) Although my scanner is exaggerating, this really is more of an orange than a pure yellow, although it does make a nice golden glaze.

The colour of joy is a mixed pigment


Over the past couple years I have primarily been focusing on watercolour as a medium, and building up my collection of watercolours one tube at a time.  I’ve been focusing on single-pigment colours, especially those rated “very lightfast” by independent testers. I now have quite a large selection of different colours, fairly well distributed around the colour wheel, yet I still felt like something was missing.

A couple of months ago, I was swatching out dot sheets by Schmincke.  I got to the colour “May Green” and squealed out loud.  It is such a lovely, happy green!  Of course, it was also a mixed pigment (PG151 and PG7) so I debated whether I should get it, since I could just mix it myself.

Around the same time, the single pigment Py117 Greenish Yellow by Holbein had caught my eye.  Although I love the colour, I was put off by Handprint’s review, which showed less-than-perfect lightfastness (Holbein ranks the lightfastness as “excellent”)

Greenish Yellow by Holbein (Py117) and May Green by Schmincke (PG7, PY151)
Two lovely yellowish greens. One has debatable lightfastness, the other is a convenience mix of two greens

I debated whether to buy these paints for a while.  I really, really wanted them, but my dogmatic side prevented me from pulling the trigger, until one day, while working on a coloured pencil piece, I noticed which pencils were worn down to stubs in each of my sets.  Without fail, it was all my yellowish greens in the light olive to bright spring green family.

Favourite greens in other media
Most used colours in other media (Neon Green by Prismacolour scanned teal for some reason, it actually is a neon yellow-green)

This should not come as a surprise.  A fun fact that few people know about me is that I have synaesthesia.  In fact, I only found out a few years ago that I have it.

Synaesthesia is a quirk in brain wiring that affects something like 3% of the population – causing different senses or concepts to be associated/linked in unusual ways.  When I first read about synaesthesia, the article used some examples that seemed really foreign and crazy to me, such as one person who has shape-smell synaesthesia (ie- cubes smell like carrots, etc), and another who uses her number-depth synaesthesia as a way to perform complex arithmetic really quickly by “balancing” distances.  I was kind of jealous that other people  had these crazy, weird superpowers when my brain was just normal (or so I thought).

The second time I read about synaesthesia was in a much more in-depth article that gave examples of more common types of synaesthesia.  This time I was very confused, because several of the examples given were things that I do experience.  For example, as I am writing this in my cold basement, my fingers feel a very nearly white, icy blue, whereas my arms and hands are more of a teal.   I have temperature-colour synaesthesia.  Since I’ve experienced temperatures as colours as long as I can remember, I have trouble imagining the concept of a warm day that is NOT orange-yellow,  or of cold fingers that are not icy blue.  I thought, since people often discuss the concept of warm vs. cool colours, and that water taps are labelled with blue/red, that this was the universal experience of temperature (although I always did find it odd that the cold tap is labelled with ultramarine, which is most definitely not the “cold” blue).

In fact, I have many types of experience/colour synaesthesia.  Among other things, I experience emotions as colours, some kinds of sounds have colour associations,  people have colour associations.


Testing Greens
“Joy” and “Self” greens in various media


Notably, my colour association for the emotion “Joy” is a bright, spring green, nearly a perfect match for Schmincke’s May Green watercolour.  You can see similar examples in the top row of the swatches of my most used/favourite colours in other media below.   No surprise then, that my reaction to painting it out was so giddy. (Again, ignore Prismacolor’s neon green.  It is neon green, not teal)

On the second row of swatches are some more olive yellow-greens.  These are all near matches for the colour I associate with myself.  The truest match for “Lee Green” is between Faber Castell Polychromos May Green, and Lyra Polycolor Apple Green.  Holbein’s Greenish Yellow is among the closest single pigment watercolours, although a little too muted and yellow.  Daniel Smith’s Serpentine Genuine (from the Primatek line) is also fairly close, but a little too cool, and not as transparent as I often want.

So, needless to say I bought both of the watercolour paints.  I will conduct independant lightfastness tests for the Holbein paint, but even if I am not satisfied it will have a place in my travel/sketch palette.  May Green may be a mixed pigment, but Joy should always be available in my palette.

I hope you will all join me in hoping for more single pigment greens to be discovered and developed for watercolour!


Effects of water hardness on watercolour paints

For the past couple of years,  my tap water has been curdling my paints.

It all started with a tube of Quinacridone Rust (PO48)  from M. Graham.   I was attracted to this paint partially because the pigment is listed as non-granulating,  and I fell in love with the colour, but ever since I got it, the paint has had a tendency to clump and curdle when mixed with water in the palette or dropped into a wash, resulting in an incredibly pronounced granulation pattern.  This pattern becomes even more pronounced when Quin. Rust is mixed with another dark valued, non granulating paint such as a phthalo.

Colour mixing experiments with Quin Rust (PO48) and Phthalo Turquoise (PB16)

I searched the internet extensively, and could find no other reports of quinacridone rust granulating like this.  For months I was actually convinced I had received a dud tube, but avoided investigating further, because I really adore the colour and transparency of the tube I own.  Instead, I simply developed strategies for mitigating the crazy clumping and curdling.

Then, a few months ago, I purchased two more quinacridone paints by M. Graham – Quinacridone Rose (PV19)  and Neutral Tint (PV19 + PG7).  Again, all of the pigments are listed as non-granulating, but both of these paints showed the same clumping/curdling behaviour (albeit to a lesser extent).

Now, one dud tube would be easy to explain, but 3 dud tubes purchased at different times would reflect very poorly on the manufacturer.  And yet, still I could find no reports of similar experiences online.  I felt like I was going crazy.

PV19 Quinacridone Rose (M. Graham) on palette with KW water

Finally, I reached out to other botanical artists and asked around.  Someone suggested I should paint with distilled water, as acidity and mineral in tap water can react with some paints.  Suddenly it clicked.   I live in Kitchener-Waterloo, known to have the hardest water in all of North America.

Water hardness refers to the amount of dissolved calcium and magnesium in the water.  Water tends to be harder when drawn as groundwater rather than from lakes or streams.  Water hardness leads to limescale deposits in pipes and fixtures, and premature degradation of water heaters in pipes.  In my area, most households (including my own) have a home water softener, which functions by replacing the dissolved calcium/magnesium with sodium.  We shower and do our dishes with saltwater, which is not drinkable, and have a hard-water tap direct from the city for drinking water. At my art studio, a few blocks away, we do not have a softener.  Thus, my paintings in the past few years have all been done with limewater, and perhaps occasionally saltwater.

In order to test the effects of the tap water on my paints, I purchased a large jug of distilled water at the grocery store, and painted out samples wet-in-wet and wet-on-dry on identical paper using hard water vs. distilled water.


The difference between the two samples was dramatic.  However, there was still a little bit of texture visible in some of the paintouts even when using distilled brushwater.  I attribute this to mineral build-ups in the pans themselves, as well on the brush and the plates I use as palettes, which of course get washed in tap water.  I even tried a sample using fresh tube paints, which was even cleaner, but still contained a little bit of granulation.

So, from now on, I will be painting using distilled water.  My paints are much posher than me, and will get fancier drinking water than I do.  Eventually, I will have a reverse osmosis system installed for the drinking water line at home, although likely not at the studio, which is a rental shared with other artists.  I will however be able to fill up a bottle at home and take it to the studio.  However, I can’t imagine going through the trouble of only washing my palettes and brushes with distilled/RO water, so some hardness effects will still be seen.

I am somewhat torn about manufacturers.  I love the brightness and saturation of M. Graham paints, but they seem to be much more prone to this kind of curdling. Since running this test, I did look through my mixing charts and found a few other paints where some minor unexpected texture, but nothing remotely close to what is seen in these three paints.

My quinacridone pigments from other manufacturers do not show nearly as much clumping.  My PV19 Permanent Rose from Winsor Newton is unfortunately not as bold and red as the M. Graham version, but it shows only the faintest amount of texture even in hard water.  My PR122 Purple Magenta and PV55 Quinacridone Purple from Schmincke and PO49 Quin. Gold from Daniel Smith are also quinacridone pigments, but paint out completely smooth.  Given that in most cases I prefer smooth, non-granulating paint, and that my paint will no doubt continue to come into contact with some dissolved calcium and magnesium, would I be better served switching to different manufacturers for quinacridones (and any other paint that may be similarly affected)?

Does anyone know what the chemical reaction occuring might be in these cases?

Schmincke Horadam Dot Sheets – Brand Review

As I mentioned in my Watercolour Palette Tour teaser post last week, I recently received some watercolour trial dot cards of Schmincke Horadam watercolours.

I do already own a few Schmincke watercolours – Purple Magenta (PV122) and Translucent Orange (PO71) are in my “greatest hits” palette.  I also keep Schmincke’s Ultramarine Finest (PB28) in my larger studio palette – although I don’t love PB28 in, it is such a ubiquitous pigment that I do want to have some around. I do find this smoother, less granulating version less aggravating than most.

In general I had been impressed with my Schmincke watercolours.  The Purple Magenta and Translucent Orange are both unique, brilliant, beautiful colours with fantastic mixing and layering properties.  All three pour buttery smooth into pans, set nicely and rewet beautifully.  However, since Schmincke is such a hard-to-find, expensive brand in Canada, I only have a few tubes that other artists had recommended as standouts from the line.   I was thrilled to see them offering trial sheets, so I could find out how the rest of the line compares.

In addition to the 12-colour card I already blogged about,  I also got a a trial sheet of the 35 new colours they released in celebration of their 135th anniversary, and a large 80-dot set.   I also snapped up some small tubes of  Cobalt Blue Deep (PB74) and their newly released Quinacridone Purple (PV55) and Phthalo Sapphire (PB15:6) in the same (135-year!) anniversary sale where I got the dot samplers.

schmincke new 35

I had hoped that that the 80-colour and 35-colour would be completely different sets (adding up to almost the full range of 139 colours Schmincke now offers), but alas there was substantial overlap, so I missed out on a number of interesting pigments, including the Translucent Brown I was particularly excited to sample.  Schmincke does produce a 135 colour dot trial sheet, which was out of stock.  Perhaps someday I will buy that one.

Schmincke testing

According to Handprint, Schmincke is a popular brand among botanical artists “for their consistent texture and less emphatic chroma”. I’ve heard the same sentiment echoed by other artists.  With only a few paints, I couldn’t really comment  on consistency, and “less emphatic chroma” is not a word I would use to describe Purple Magenta or Transparent Orange, even alongside my eye-scorchingly bright M. Grahams.  After painting out swatches from the large trial sheet (and squeezing my new paints into pans), I have a much better understanding of what they mean.

Schmincke Horadam are exceptionally well-behaved paints.  These are stern, precise, German paints.  

All 6 of the tubes I own have squeezed out neatly with a creamy, even consistency, and then settled into looking like smooth, shiny little candies, which all rewet neatly and are never sticky.  This in comparison to my M. Graham paints- which glob out unevenly into sticky, unsightly mounds in my palette which eventually dry but never look smooth, or my Daniel Smith and Winsor Newton paints, which crack as they dry or pop out of the pan entirely.

In terms of the behaviour of the paints themselves, particularly in the dot set, I was struck with how all of the paints behaved similarly.  Whereas in some brands, like Daniel Smith, granulating paints will create wild crater-like patterns and staining paints will be prone to dramatic backruns, this effect was barely noticeable across most of the Schmincke line, with the exception of a few of the newest additions.  Granulating paints left delicate granules in place.  Staining paints painted out evenly, and stayed put where they were brushed in.   The colours seemed about as saturated as Winsor and Newton or Holbein overall, although the range of saturation was less pronounced (the phthalos less overwhelming, cobalts and earth tones more saturated).

The Schmincke line does include a few very bright, brilliant, high chroma colours.However, although these are all bright and well-pigmented, they do not overwhelm mixes as much as some of my M. Grahams or even the Winsor Newton phthalos.

Having swatched out these paints, I can certainly see why many botanical artists would appreciate this controlled, predictable behaviour.   In very detailed, technical work where precise colour mixing is desired, one does not want to worry about a stray puddle of phthalo green eating your palette whole or crazy textured bumps appearing in your smooth flowers because your ultramarine granulated over-enthusiastically.  Schmincke definitely delivers a consistent, reliable, user-friendly product.

For most of my painting, I am addicted to the wild ride of the most saturated, highly staining, brightest colours and the responsiveness of honey-based paints.  However, I do find the idea of putting together a cohesive palette with similar hues to my current favourites, but all in nicely-setting, evenly mixing formulations appealing.  Occasionally for delicate subjects, or in a travel context, it might be nice to have better-behaved alternatives, so I looked carefully at the Schmincke offerings of my favourite pigments.

My current primary triad is Schminke’s Purple Magenta (PR122),  M. Graham Nickel Azo Yellow (PY150) and Winsor Newton Pthalo Turquoise (PB16).  Given that my choice for magenta is already Schmincke, I was particularly interested in Schmincke’s PY150 and PB16 offerings.  I was not impressed with either one.  Whereas the M. Graham PY 150 looks like bird poop in the pan but washes out to a brilliant primary yellow, the Schmincke looked smoother concentrated but didn’t achieve that clear pure yellow look even in tints- always looking a little brownish/earthy, not a great choice for a primary/main yellow.  And while the PB16 was a lovely colour in tints, it did not appear to be as concentrated as the Winsor Newton, which is nearly black in masstone.

Other favourites were also dissapointing.  PY129 did not have the lively green masstone of M. Graham’s green gold, instead appearing more brown.  PB60 Delft Blue was the gorgeous blue of cobalt glass when wet, and I was very hopeful, but when dry it looked ashy and greyer than both the M. Graham and Daniel Smith versions.

However, the dot cards did introduce me to some other offerings which I may soon add to my collection.  Schmincke has quite a range of pleasingly bright and transparent colours in the red/rose/coral family. I am also hungrily eyeing the absolutely yummy version of PG50 Cobalt Teal.  I currently own this pigment in QoR, which is nice from the tube but I don’t like the rewetting behaviour of.  The Schmincke version is prettier, brighter, and rewets beautifully.   I may even spring for May Green, which is the happiest bright green (gasp! a convenience colour).