There are 28 collagen types, and three dominate: type I, the main collagen of tendon and hide; type II, the collagen of cartilage; and type III, which sits alongside type I in hide and is a major part of large blood vessels. Once a supplement is hydrolysed into peptides, the type number mostly names the tissue it was made from. Undenatured type II is a different product, sold by the milligram.
Type numbers are printed on collagen labels as though they were grades, and the question “which type is best” follows naturally. The honest answer depends on one fact the type number does not tell you: whether the collagen has been cut into peptides or kept whole. Everything below is anatomy and processing. None of it is a claim about what any type does for a person.
How many types of collagen are there?
Twenty-eight. Ricard-Blum’s 2011 review in Cold Spring Harb Perspect Biol counts 28 members of the collagen family, each containing at least one triple-helical stretch. Several are confined to particular tissues. Types I, II and III are the major structural collagens, with long, uninterrupted triple helices built on a repeat in which every third amino acid is glycine (Krane, Amino Acids 2008). They are the three a supplement label is likely to name.
The triple helix is what makes a collagen a collagen: three chains wound around a common axis, held in shape by a high content of proline and hydroxyproline. The types differ in which chains they use and where the body puts them. Type III, for instance, is made of three identical chains (Kuivaniemi and Tromp, Gene 2019), while the type II fibril in cartilage carries type IX molecules bonded to its surface and type XI forming its core template (Eyre and colleagues, Eur Cell Mater 2006).
What is the difference between type 1, type 2 and type 3 collagen?
Where each sits and how its chains are built. Type I is the main collagen of tendon, where collagen makes up 65 to 80% of the dry mass (Kannus, Scand J Med Sci Sports 2000). Type II is the collagen of cartilage, which is two-thirds collagen by dry weight (Eyre and colleagues, Eur Cell Mater 2006). Type III is built from three identical chains and is a major structural part of large blood vessels and the bowel (Kuivaniemi and Tromp, Gene 2019).
| Type | Where in the body | Common supplement source | How it is sold |
|---|---|---|---|
| Type I | Most of the collagen in tendon; most of the collagen in hide | Cattle hide, pig hide, fish | Hydrolysed peptides or gelatin, grams per serving |
| Type II | Cartilage, where collagen is two-thirds of the dry weight | Cartilage, such as chicken sternum | Hydrolysed, in grams; or undenatured, at about 40 mg |
| Type III | Alongside type I in hide; large blood vessels and bowel | Arrives with type I in hide | As part of hide collagen, printed “Type I and III” |
| The other 25 | Several confined to particular tissues; types IX and XI sit on the type II fibril | None as such | Not a supplement category |
The source column is the one a buyer can use. Hide and fish give mostly type I; cartilage gives type II; hide brings a share of type III with it. The hide figure has a number on it: about 13% type III in one analysis of insoluble cattle hide collagen, the rest type I (Scott and Veis, Connect Tissue Res 1976). The same raw-material logic, cattle against fish, is set out in bovine vs marine collagen.
Does the collagen type matter once it is hydrolysed?
Much less than the label suggests. Hydrolysis destroys the triple helix and cuts each chain into short fragments sharing the same glycine, proline and hydroxyproline chemistry. When Iwai and colleagues gave people hydrolysates made from pig material, chicken feet and cartilage, collagen fragments in their blood peaked one to two hours later, and the main one was the same, proline-hydroxyproline (J Agric Food Chem 2005). On a hydrolysed product, the type number mostly names the source tissue.
The word “mostly” is doing honest work. The types do differ in sequence, and a cartilage hydrolysate is not chemically identical to a hide hydrolysate. What has not been shown is that the difference survives digestion in a way that matters. The fragment that dominates the blood after a hide-based, a foot-based and a cartilage-based hydrolysate is the same two-amino-acid piece, and that finding comes from blood samples, not from any tissue.
Which type of collagen is best?
There is no best type in general, because the types do different jobs in the body and, once hydrolysed, arrive in the blood as very similar fragments. The useful questions are narrower: is the product hydrolysed or undenatured, what tissue was it made from, and how much is in a serving. For the tendon question this site covers, the trials used gelatin or collagen peptides rather than cartilage products, and they disagree with each other.
The disagreement is the part usually left out. The study most dose advice traces to, Shaw and colleagues, Am J Clin Nutr 2017, compared 5 g and 15 g of vitamin-C-enriched gelatin in eight healthy men and measured a blood marker, not a tendon. Balshaw and colleagues, Med Sci Sports Exerc 2023, randomised 39 men to 15 g of collagen peptides or placebo a day for 15 weeks of training, imaged the tendon by MRI, and found no between-group difference. The full ledger is in the collagen research summary, and the dose arithmetic in how much collagen for tendons.
What is undenatured type II collagen?
A different product that shares a name. It is type II collagen processed so that its native structure survives, sold at about 40 mg a day, the dose in a 120-day trial that used material from chicken sternum (Lugo and colleagues, J Int Soc Sports Nutr 2013). Its proposed mechanism is oral tolerance, an immune response said to depend on structures present in undenatured collagen but not in hydrolysed collagen (Harris and colleagues, J Diet Suppl 2022).
Three things follow, and none of them is a statement about whether it works. First, the dose: 15 g of hydrolysed collagen is 375 times the mass of 40 mg, so the two are not competing for the same job; they are not even measured on the same scale. Second, the evidence: undenatured type II has its own literature, run on its own outcomes, and nothing in the hydrolysed-collagen research transfers to it or from it. This site does not review that literature and makes no claim about it.
Third, and most usable, the label. Harris and colleagues examined type II powders and finished capsules on the market and found that powders labelled as undenatured type II had markedly different properties, from fibre size under the electron microscope to the structural signal their assay was built to detect. Their advice is not to assume that products sold under that name are interchangeable. It is a laboratory characterisation, not a trial in people, but it makes the useful point: the word on the front describes processing, and it can be more or less true of what is inside.
What does “Type I and III” on a collagen label mean?
That the collagen was made from a tissue that contains both, which for a bovine powder usually means hide. Cattle hide collagen is mostly type I with a minority of type III: about 13% type III in one analysis of insoluble hide collagen (Scott and Veis, Connect Tissue Res 1976). The phrase describes the raw material rather than a blend of separate ingredients, and no study cited here links the share of type III in a powder to anything measured in a person.
Read that way, a type line is a sourcing statement, and it can be checked against the rest of the label: “bovine” and “Type I and III” fit together, as do “marine” and “Type I”, or “chicken sternum” and “Type II”. A type line that does not fit its stated source is worth a question to the seller. The other terms on the same panel are defined in the supplement label glossary.
What the research does not show
- A type comparison on any tissue. We found no trial comparing hydrolysed type I with hydrolysed type II on a tendon, cartilage or any other tissue measurement.
- That fragments in blood reach a target. The shared fragment after pig, chicken-foot and cartilage hydrolysates is a blood finding, measured one to two hours after a single dose.
- Anything about undenatured type II here. It is dosed at a 375th of the mass, is proposed to act through a mechanism that is different in kind, and has a separate literature that this site does not evaluate. Its products also differ from each other on laboratory measures.
- A settled tendon answer for any type. The trials behind the dose used gelatin and peptides, the best-known of them measured a blood marker in eight men, and the 15-week MRI trial at 15 g found no difference from placebo.
Where Under Load sits
Under Load is bovine collagen peptides, Type I and III: one ingredient, 350 g, hydrolysed rather than undenatured, so by this page’s reading its type line names a raw material and implies no separate mechanism. The declared serving is 10 g, two 5 g scoops, 35 servings; the protocol on this site is a directions statement, 15 g, three scoops before training. Its certificate was run by ViaCore, the manufacturer’s own laboratory, on lot 121125450, and gives protein as 9.8 g measured per 10 g serving; 1.5 servings is 14.7 g by arithmetic. What that document covers is transcribed on Testing & Certification.
More in this cluster: collagen peptides vs gelatin, collagen vs bone broth, how to choose a collagen supplement and every supplement form compared.
Sources
- Ricard-Blum S. Cold Spring Harb Perspect Biol 2011;3(1):a004978. PubMed 21421911
- Krane SM. Amino Acids 2008;35(4):703-710. PubMed 18431533
- Kannus P. Scand J Med Sci Sports 2000;10(6):312-320. PubMed 11085557
- Eyre DR, Weis MA, Wu JJ. Eur Cell Mater 2006;12:57-63. PubMed 17083085
- Kuivaniemi H, Tromp G. Gene 2019;707:151-171. PubMed 31075413
- Scott PG, Veis A. Connect Tissue Res 1976;4(2):117-129. PubMed 131673
- Iwai K, Hasegawa T, Taguchi Y, et al. J Agric Food Chem 2005;53(16):6531-6536. PubMed 16076145
- Lugo JP, Saiyed ZM, Lau FC, et al. J Int Soc Sports Nutr 2013;10(1):48. PubMed 24153020
- Harris RB, Fonseca FLA, Sharp MH, Ottinger CR. J Diet Suppl 2022;19(6):717-732. PubMed 34060435
- Shaw G, Lee-Barthel A, Ross ML, et al. Am J Clin Nutr 2017;105(1):136-143. PubMed 27852613
- Balshaw TG, Funnell MP, McDermott EJ, et al. Med Sci Sports Exerc 2023;55(11):2083-2095. PubMed 37436929